stack overflow with RTAI 3.1 - test4 on 2.6.7

David Rolenc drolenc at
Fri Oct 1 17:06:39 CEST 2004

A modification to minimize propagation of interrupts to linux in my 
driver fixed, or made less likely, the problem. I still am not convinced 
in the validity of this solution, but have not seen the problem in days. 
Thanks for everone's assistance.


David Rolenc wrote:

> Paolo,
> I have tried on a single CPU with similar results, although I still 
> used the rtai_smp module and rebooted linux with maxcpus=1 on the 
> command line. Not really a recompile. I have taken the same software 
> and put it on another similarly configured box, just different 
> hardware (Dell 2650). It works fine there. I think this may be a 
> hardware issue with Dell 2600s. They use an E7501 shipset on the 2600 
> , while they user a ServerWorks GCLE on the 2650. Has anyone had 
> issues, aside from SMI , with the E7501s?
> Thanks,
> -Dave
> Paolo Mantegazza wrote:
>> David Rolenc wrote:
>>> I have a Dell Poweredge 2600 System with dual 2.4 Xeons, 2 Gb of 
>>> Ram, HT disabled at the BIOS, swap disabled and SMI interrupts 
>>> disabled. I have a piece of custom hardware that I am using to pull 
>>> in data. I have installed a real-time interrupt handler for this 
>>> (~6000 interrupts/sec). The handler basically starts a DMA and 
>>> determines whether a wake up of other tasks is necessary. In the 
>>> event that a wakeup is needed, the rt interrupt handler propagates 
>>> the interrupt to linux, at which time a regular linux wake_up() 
>>> occurs. I should mention that during this time there is heavy disk 
>>> I/O, as I am archiving and reading the data in question. The only 
>>> realtime task is the interrupt handler mentioned. I am running into 
>>> stack problems that seem to be related to interrupts. 
>> Apart from my suggestion to try on an ADEOS only test, could you see 
>> how what above works by compiling Linxu and RTAI with just one 
>> processor?
>> Paolo.
> _______________________________________________
> RTAI mailing list
> RTAI at

More information about the Rtai mailing list