Embedded Software for SoC - Grupo de Mecatrônica EESC/USP

18 Chapter 2
Table 2-1. Worst case single process analysis and measurements.
Process
Measured
WCET in
scratchpad
Calculated
max # of
cache misses
Calculated
WCET
w/ cache
Measured
WCET
w/ cache
Measured
WCET
w/o cache
a
b
c
79
79
104
the worst case number of cache misses. The third column contains the worst
case execution times from external memory with cache calculated using the
SYMPTA/P approach. The measurements from external memory – with and
without cache – are given in the last two columns.
5.2. RTOS analysis
Apart from influencing the timing of individual tasks through scheduling‚
the RTOS itself consumes processor time. Typical RTOS primitive functions
are described e.g. in [1]. The most important are: task or context switching
including start‚ preemption‚ resumption and termination of tasks; and general
OS overhead‚ including periodic timer interrupts and some house-keeping
functions. For formal timing analysis to be accurate‚ the influence of RTOS
primitives needs to be considered in a conservative way.
On the one hand‚ execution time intervals for each RTOS primitive need
to be considered‚ and their dependency on the number of tasks scheduled by
the RTOS. The second interesting question concerns patterns in the execution
of RTOS primitives‚ in order to derive the worst and best case RTOS
overhead for task response times.
Ideally‚ this information would be provided by the RTOS vendor‚ who has
detailed knowledge about the internal behavior of the RTOS‚ allowing it to
perform appropriate analyses that cover all corner cases. However‚ it is
virtually impossible to provide numbers for all combinations of targets‚ compilers‚
libraries‚ etc. Alternatively‚ the RTOS vendor could provide test patterns
that the integrator can run on its own target and in its own development
environment to obtain the required worst and best case values. Some OS
vendors have taken a step in that direction‚ e.g. [11].
In our case‚ we did not have sufficient information available. We therefore
had to come up with our own tests to measure the influence of ERCOSEK
primitives. This is not ideal‚ since it is tedious work and does not guarantee
corner-case coverage. We performed our measurements by first instrumenting
accessible ERCOSEK primitives‚ and then using the LSA-based approach
described in Section 5.1. Fortunately‚ ESCAPE (Section 4.1) generates the
ERCOSEK configuration functions in C which then call the corresponding
ERCOSEK functions (object code library). The C functions provide hooks for
instrumentation.