On Fri, 24 Aug 2018 at 15:39, john zhang laojianghusz@163.com wrote:

Once job test case make device cpu crash, is there any mechanism by choose to restart next round test or end test?

Currently, the test job must simply timeout and fail, so the test will end. LAVA spends a lot of time parsing the output of the DUT - if that output stops, LAVA cannot assume that the kernel has crashed and not all kernel crashes produce recognisable messages, so the only option is a timeout.

Support for moving on to the next test: action after specific timeouts is in development.

https://review.linaro.org/#/c/lava/lava/+/26442/ and https://projects.linaro.org/browse/LAVA-1371