Re: [Xenomai-help] keyboard interrupt

[Jan Kiszka <jan.kiszka@domain.hid>]

[-- Attachment #1: Type: text/plain, Size: 1549 bytes --]

Harco Kuppens wrote:
> Hi,
> 
> I tried to make a demo program which showed how that user space task
> waiting for an interrupt with rt_intr_wait automatically gets an
> increase in priority when the interrupt arrives so that it can be dealed
> with immediately.
> 
> In my program I use the keyboard interrupt IRQ=1. I have a priority task
> doing spinning for 3 seconds using 100% of theCPU. During this spinning
> I want to test if a low priority task can still get an interrupt from
> the keyboard by letting the end user type some letters on the keyboard when
> My program works but only  the first interrupt gets caught when the high
> priority task is running. The following interrupts are only dealed with
> when the high priority task is done.
> Probably because linux first has to finish the first interrupt?

Exactly. You forward the IRQ to Linux, thus you reply on Linux for
dealing with the periphery (acking the IRQ there) and the final
end-of-IRQ signal.

> I thought that the adeos pipeline kept  interrupts on a hold between
> xenomai and linux, until xenomai has done its work. But in my example it
> seems that the second interrupt has to wait for linux has finished
> handling the first interrupt. That is really something you don't want.
> 
> Is my interpretation wrong, or do I something wrong in my code?

IRQ sharing between deterministic and non-deterministic code (here:
Xenomai and vanilla Linux) will never work, that's an inherent design
issue, nothing Xenomai or ipipe-specific.

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 258 bytes --]