Summary: Coping With Interrupt Execution Time in RT Kernels: a
ReTiS Lab, Scuola Superiore S. Anna, Pisa
This paper describes an on-going work focused on using Adaptive Reservations to cope
with unpredictabilities due to the interrupt processing times in real-time kernels. Using
this approach, the execution time stolen by interrupt processing is modeled as a variation
in the application execution times, and an adaptive scheduling mechanism automatically
compensates its eects. The proposed approach has been implemented in Linux/RK, and
some preliminary experimental results are presented.
In the last years, there has been a growth of interest in running applications with real-time
requirements (such as streaming applications, software mixers, and so on) on top of desktop
Operating Systems (OSs). For this reason, a lot of eort has been spent in providing general
purpose desktop OSs (such as Linux, FreeBSD, or Windows) with real-time capabilities.
However, omitting some notable exceptions, these works mainly focused in implementing
real-time scheduling algorithms for the CPU and other system resources inside the OS kernel.
Most of these algorithms are based on some very stringent assumptions, such as the full