The software unit verification plan covers the content of the work products to be verified, the verification
tools,methods and environment with pass and fail criteria. This task is carried out according to ISO 26262, vol 8,
clause 9 where the test object described relates to the software units in this context.
For model based development the corresponding parts of the implementation model represent test objects as they
implement the software units.
The intent is that the software units tests shall demonstrate that the software units should be robust with enough
resources to support the functionality and achieve compliance with the software unit design specification, and the
hardware software interface. The test should also verify that the software not only has the correct
implementation of the intended functionality but also an absence of unintended functionality.
Methods for Software Unit Testing
This table/figure is taken from ISO 26262:2011-9.4.3 © ISO 2011 – All rights reserved. Reproduced with the
permission of ANSI on behalf of ISO.
The table above indicates the methods the test should take depending upon ASIL, it should be noted that the
tests should comprise a combination of methods as no one method is sufficient by itself.
In order to find the most appropriate method to derive test cases according to the unit test methods please refer
to "Table 11 - Methods for deriving test cases for software unit testing", located in ISO 26262, vol 6, clause 9.4.4.
This table covers the following methods with respect to the recommended ASILs
Analysis of requirements is highly recommended for items of all ASIL levels
Generation and analysis of equivalence classes is highly recommended for items of ASIL B,C or D and
recommended for items of ASIL A
Analysis of boundary values is highly recommended for items of ASIL B,C or D and recommended for items of ASIL
A , this applies to interfaces and values crossing boundaries or ranges.
Error guessing is recommended for items of all ASIL levels, these tests can be based upon data collected through
experience and/or expert judgment
It is expected that the test environment shall correspond as far as possible to the target environment. If they are
differences between the test and target environment then further tests need to be defined (based upon the differences
between the source and object codes and the different environments) to provide coverage of the unit on the target
environment.
For further information relating to this requirement the user is referred to ISO 26262, vol 6, clause 9.4.5
Copyright ISO 2011. All rights reserved. Reproduced with permission of ANSI.
|