Note it : The gcno files

Jul 15, 2018 • Vijay K. Banerjee (thelunatic)

Hello everyone!

In my previous post I have given the basic outline of the workflow that we have planned to follow for generating the gcov coverage reports from covoar. The plan is interesting but certainly not that simple, there are few challenges in this project which are a bit difficult to tackle down.

The very first of the challenges is the mystery of the GCNO and the GCDA files. According to plan, the covoar tool will read the gcno files along with the qemu trace to generate the gcno files. But… The structure of the gcno file is still a mystery to us and without solving that mystery, it is not possible to go ahead with the gcov support without understanding the format of the file we are working with.

To deal with this we have planned to create a GCNO_DUMPER program to dump the contents of the gcno files into a human readeable txt format. I have already started the dumper project and it can be followed in the following repository.

https://github.com/thelunatic/gcno_dumper

It already produces some visible and understandable txt file, the work is in progress and will soon (hopefully) be ready to dump the contents of any gcno file.

In the next post I will describe each fields in the txt file that is generated so far that will demystify the gcno mystery to some extent.

Thank you for giving it a read! :)