Introducing the "Kernel Fault Injection Framework" generation and testing capability in LTP. This is in line with the recent proposal made through LTP Paper at OLS 2009: "Putting LTP to test - Validating both the Linux kernel and Test-cases" (http://ltp.sourceforge.net/documentation/technical_papers/Putting_LTP_to_Test.pdf). This infrasturcture will help LTP directly in the following ways:
1) Allow test developers to test their new test cases against a faulted kernel, and then compare it on stable kernel run, impacting in better test development,
2) Allow test engineers to be able to generate more code coverage by traversing the rarely touched parts of the kernel code,
As we move forward in using this, we would definitely find some other advantages of this framework. This is V2 of patchset after incorporating comments from Mike, Paul & Garret.
[PATCH v2 02/05] Add Script which would actually do the job of injecting faults: Script which would actually do the job of injecting faults by changing various parametrs available under /debug/fail*. This would be done dynamically during LTP run, and, is capable of taking parameters from 0 to 100 to vary the probability of Fault the user wants to inject in the running kernel. Signed-off-by: Subrata Modak<subrata@linux.vnet.ibm.com>.
1 file changed