On Thu, Dec 8, 2011 at 10:40 PM, Sree kumar <sreekumar.c@samsung.com> wrote:
Thanks a lot for the update.
 
We are interested in the number of test cases and the quality of the test cases.
 
1.Can LAVA detect critical bugs related to UMM?
The Multimedia Working Group is currently working on some tests for UMM on their CMA test.  We'll be looking at getting some regular test runs with that when it's complete.  More information on that is available here: https://wiki.linaro.org/OfficeofCTO/MemoryManagement/Validation/CMATest
 
2.Is it useful in testing power management?
For power management we have the ability to run tests that don't require any specific
instrumentation, but we do not yet have devices in the lab to allow advanced power
consumption measurements in our lava lab instance.  Once we have such devices, we can take a look at the board level modifications needed to connect them, and the extensions in LAVA that will be needed to make use of them. We have been working with the Powermanagement team to
find a suitable solution for that payload that matches our needs in scalability and usability in the lab.
 
Do you have the details of test coverage? If “yes”, please share.

Kernel- how many test cases? which area is covered?

Based on the good feedback you gave us during connect, the kernel WG has started to
look into this more thoroughly on the kernel test front, as have the landing teams, but
they will not yet have those tests in the test pool before end of December.

LAVA is the validation *infrastructure* to automatically deploy Linux images to Linaro supported development boards, run tests on them, and collect the results into a human readable dashboard system.  For the actual tests, we *strongly* encourage contributions from members and working groups so that we can ensure their components and devices are well-tested under LAVA.  Additionally, we plan to make test cases a focus topic at the next Linaro Connect, and will do what we can to continue to pick up new tests that we know will be generally useful.

Thanks,
Paul Larson