Re: sja1000 interrupt problem

[Wolfgang Grandegger <wg@grandegger.com>]

On 11/17/2013 03:27 PM, Oliver Hartkopp wrote:
> 
> 
> On 17.11.2013 09:18, Wolfgang Grandegger wrote:
>>
>> On Sat, 16 Nov 2013 22:42:10 +0100, Oliver Hartkopp
> 
> 
>>> Btw. I would try two more things with the sja1000.c driver:
>>
>>>
>>
>>> 1. Print the corresponding CAN device name and the point in the code
>>
>> before
>>
>>> returning IRQ_NONE to catch the problematic return site.
>>
>>>
>>
>>> 2. Replace all IRQ_NONE with IRQ_HANDLED for a test ...
>>
>>>
>>
>>> Any other ideas?
>>
>>
>>
>> A function trace is still the first thing to take. Be aware that the
>>
>> tasks are running on more than one CPU. It's just to establish a good
>>
>> trigger.
>>
> 
> Yes. But so far the function trace only showed the fact of simultaneous irq
> treads for the two CAN interfaces - what we were expecting :-)

Yes, because the trigger was not good so far.

> IMO it's interesting to get the concrete return site from where the IRQ_NONE
> is returned. Maybe one of the other interfaces / bits in PITA are enabled
> which we did not see so far.

The problem is that the IRQs are shared, also with the ATA disk. There
it's normal to see unhandled interrupts. My idea was to trigger a
"tracing_off" after 10 IRQ_NONE in sequence.

>> Anyway, I think the direct "return IRQ_NONE" are wrong because they
>>
>> do not call the "irq_post" handler.
>>
> 
> That's true. I just sent a patch to fix this issue.

Thanks, will have a closer look tomorrow.

Wolfgang.