Hi Guenter, all,
I don't really understand this focus on fixing theoretic/irrelevant
race conditions in drivers which no one uses anymore. Maybe someone
can enlighten me ?
In conjunction with linuxtesting.org and The Linux Foundation, I've been
enlisted to test and work on helping to test tools they use for
reliability testing of linux-based systems. In particular, two tools,
RaceHound (whose command is 'lines2insns' and which isolates race
conditions in kernel modules) and Klever, which is browser-based, are
under continual development by the Center and I aim to help them improve
their throughput by aiding in investigating where, in the automated nature
particularly of Klever (requiring considerable configuration as well),
there may areas to improve.
Hence, yes, a large amount of our findings result in manifesting as
theoretical and possible only, but relevant to improving the tools
nonetheless.
Hope that helps with the 'enlightening' :), regards,
Mark
