[LAU] rtirq configuration with external USB DAC?

Chris Caudle chris at chriscaudle.org
Tue Aug 13 21:39:07 CEST 2019

On Tue, August 13, 2019 10:11 am, al3xu5 / dotcommon wrote:
>> The timer "rtc" and the keyboard "i8042" should be removed from the
>> list, years back they where added, but today this isn't done anymore.
> Can you tell me why they should be removed now, please?

The RT prioritization should be based on what devices and software should
always be serviced in a timely manner.  In the context of this audio group
that would mean the audio device driver(s), jackd if that is used, and
then anything else which might be critical.  The jackd server will set
priority independently, so the only entries which need to be in rtirq are
the hardware device drivers which need realtime priority.

There is no current software that I am aware of which could require
realtime access to the hardware realtime clock (despite the name it is not
related to realtime behavior in this context). Including that device in
the rtirq list just wastes processor time servicing a device with no use
for audio.

The i8042 is the emulated legacy keyboard controller.  I do not think it
is necessary, but I could accept the argument that you don't want to miss
any user interaction. It should definitely be last in the list, so the
i8042 is at the lowest realtime priority. Probably as a practical concern
there is no reason to keep i8042 in the list, user interaction is slow
enough that unless the RT software is not behaving correctly the key input
should still be serviced, and if the RT software is behaving in a way
which uses too much processor time it will probably not respond properly
to user key presses, so in either case having the keyboard controller
driver serviced with realtime priority does not seem to offer benefit.

Chris Caudle

More information about the Linux-audio-user mailing list