Spurious IRQ7

[Grigori Khmyrov]

Originally it is square wave (0.1 ms ON and 4.9 ms OFF) signal.
When this extra interrupt happens (once out of 10000 times) I'm producing
some special HW signal to synchronize the oscilloscope and check the 
original
signal. It is very clear nothing (zero level signal) I can see. Actually 
from interrupt
I also can check the original signal level. And it is also zero.

One additional things. I'm transferring about 4kB data over the network 
after
every interrupt. If I'm disabling data transfer, no extra interrupts 
happens.

Thank you for suggestions,
Grigori

Angelo Fraietta wrote:

>
>
> Grigori Khmyrov wrote:
>
>> I have problem with extra interrupts at i686 industrial computer.
>> It never happened before with old VIA C3 400 MHz (100x4.0) board.
>> But I have an extra interrupts with new one VIA C3 667 MHz (133x5.0).
>> We have a hardware which produces signal every 5 ms. We use the 
>> signal as IRQ7.
>> Very rear I see an extra call of interrupt handler. Spurious IRQ7 
>> happens. I'm sure
>> that our hardware is not producing that signal. Is it any suggestion 
>> what can be
>> wrong. Can it be 8259A Interrupt Controller or RTEMS?
>> I'm using RTEMS 4.6.2 with i486 BSP.
>> Any suggestion how to fix or catch a problem appreciated.
>>
> What does the signal look like with a CRO?
> You might be getting reflections back that are causing reflective 
> spikes (this happened to me and appeared as extra pulses) so check 
> your impedances