How interrupts are handled in RTEMS

Sebastian Huber sebastian.huber at embedded-brains.de
Wed Jul 26 13:19:48 UTC 2017


On 26/07/17 15:01, Denis Obrezkov wrote:

> Hello all,
>
> I have read RTEMS manuals but still can't fully  understand how interrupts
> are handled.

The manuals will be of no help here.

>
> For example, when an interrupt occurs pc is set to an IRQ handler address.
> So, the program's execution continued from that place.
> And my question is: how do set_vector and rtems_interrupt_catch relate
> to each other and to that IRQ function address installed in a processor
> during initialization?

The set_vector() and rtems_interrupt_catch() are legacy functions. They 
should not be used for new ports. The set_vector() is not documented. 
The rtems_interrupt_catch() lacks important features like an argument 
pointer for the handler.

For RISC like machines the ARM

cpukit/score/cpu/arm/arm_exc_interrupt.S

and PowerPC

c/src/lib/libcpu/powerpc/new-exceptions/bspsupport/ppc_exc_async_normal.S

sources are a good reference. The SPARC is too complicated due to its 
register windows. The ARM is also a bit complicated due to the different 
processor modes. I would start with a look at the PowerPC code.

In general you do the following:

Save volatile registers
Increment thread dispatch and ISR nest level
Switch to interrupt stack if necessary
Call high level handler (installed via rtems_interrupt_handler_install())
Decrement thread dispatch and ISR nest level
Do a thread dispatch if necessary and allowed
Restore volatile registers
Return from interrupt

-- 
Sebastian Huber, embedded brains GmbH

Address : Dornierstr. 4, D-82178 Puchheim, Germany
Phone   : +49 89 189 47 41-16
Fax     : +49 89 189 47 41-09
E-Mail  : sebastian.huber at embedded-brains.de
PGP     : Public key available on request.

Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG.



More information about the devel mailing list