[LAU] hdsp not happy with recent kernels

[Fernando Lopez-Lezcano]

On 09/15/2012 01:25 PM, peter at peterlutek.com wrote:
>> did some observing...
>> when i put snd_hdsp in the RTIRQ_NAME_LIST in the rtirq config, neither
> priorities nor IRQs change. however...
>> for ALL kernels under which jack fails:
>> /etc/init.d/rtirq status
>> shows irq/19-snd_hdsp AND irq/19-firewire
>> and for the one kernel under which jack works:
>> /etc/init.d/rtirq status
>> shows irq/19-snd_hdsp and a DIFFERENT irq for firewire
>> this seems significant.
>> and... how can i ensure firewire gets a different irq?
>> for now, i'll try killing the  firewire module, since i don't use
> anything
>> firewire at the moment.
>> cheers!
>> .pltk.
>
> experiment results:
>
> on a 3.5-3 kernel, after killing firewire modules, rtirq status is:
>
>    PID CLS RTPRIO  NI PRI %CPU STAT COMMAND
>     45 FF      90   - 130  0.0 S    irq/8-rtc0
>    757 FF      85   - 125  0.0 S    irq/19-snd_hdsp
>    767 FF      80   - 120  0.0 S    irq/44-snd_hda_
>    235 FF      75   - 115  0.0 S    irq/16-ehci_hcd
>    236 FF      74   - 114  0.0 S    irq/23-ehci_hcd
>     44 FF      70   - 110  0.0 S    irq/1-i8042
>     43 FF      69   - 109  0.1 S    irq/12-i8042
>     32 FF      50   -  90  0.0 S    irq/9-acpi
>    233 FF      50   -  90  0.0 S    irq/16-mmc0
>    239 FF      50   -  90  0.8 S    irq/41-ahci
>    641 FF      50   -  90  0.0 S    irq/42-mei
>    708 FF      50   -  90  0.0 S    irq/17-rtlwifi
>    716 FF      50   -  90  0.0 S    irq/43-i915
>   2371 FF      50   -  90  0.0 S    irq/40-eth0
>      3 TS       -   0  19  0.2 S    ksoftirqd/0
>     15 TS       -   0  19  0.1 S    ksoftirqd/1
>     19 TS       -   0  19  0.1 S    ksoftirqd/2
>     23 TS       -   0  19  0.1 S    ksoftirqd/3
>
> ...and after my script which kills other unneeded things:
>
>    PID CLS RTPRIO  NI PRI %CPU STAT COMMAND
>     45 FF      90   - 130  0.0 S    irq/8-rtc0
>    757 FF      85   - 125  0.0 S    irq/19-snd_hdsp
>    235 FF      75   - 115  0.0 S    irq/16-ehci_hcd
>    236 FF      74   - 114  0.0 S    irq/23-ehci_hcd
>     44 FF      70   - 110  0.0 S    irq/1-i8042
>     43 FF      69   - 109  0.1 S    irq/12-i8042
>     32 FF      50   -  90  0.0 S    irq/9-acpi
>    233 FF      50   -  90  0.0 S    irq/16-mmc0
>    239 FF      50   -  90  0.5 S    irq/41-ahci
>    641 FF      50   -  90  0.0 S    irq/42-mei
>    716 FF      50   -  90  0.0 S    irq/43-i915
>   2371 FF      50   -  90  0.0 S    irq/40-eth0
>      3 TS       -   0  19  0.1 S    ksoftirqd/0
>     15 TS       -   0  19  0.1 S    ksoftirqd/1
>     19 TS       -   0  19  0.1 S    ksoftirqd/2
>     23 TS       -   0  19  0.0 R    ksoftirqd/3
>
> ...then, jack still fails. same story for the 3.2.0 kernel.
>
> so, avoiding the irq conflict by unloading firewire modules doesn't solve
> the problem. mrph.

Argh, that seemed like a good lead...
What does "cat /proc/interrupts" show? I think the interrupt count for 
each cpu should increment. Does it?

Sorry, don't quite know where to go from here, looks like something for 
kernel gurus in lkml...
-- Fernando