While doing the fs_mark measurements, and with ext4 in 2.6.28, I
looked back at what a test plan would be for ext4.
I tried again gcov with the help of Peter (I was hoping that it would
become mainline with 2.6.28).
Here are the results:
While I do not share the idea that 100% coverage is a requirement, I
found it useful the show obvious holes in a test suite and as an
indication for the test effort.