Martin Doucha

Christian Amann has rewritten the fstat tests back in June. Is this issue still relevant?

There is also a socket hang regression related to the CVE bugfix which should be tested. See [packetdrill script](https://bugzilla.suse.com/attachment.cgi?id=810193) that can reproduce the issue on incorrectly patched kernels (e.g. 4.12.14-197.7.1...

It'd make more sense to let `tst_brk(TCONF)` skip only one test case instead of terminating the whole test process. Test cases are often not interdependent. But that would require one...

It's the simplest option for setting up the test case run but cleanup will be much more difficult than with pthreads. Unless you want to ignore cleanup entirely and terminate...

And there's also the question whether we want `tst_brk(TBROK)` to skip one test case or terminate the whole test. I'd vote for terminating everything, which would require extra code in...

After careful review of the linked commit, the CVE appears to be bogus. Both functions that call `valid_master_desc()` append an extra null byte to `new_desc` beforehand, which means that `memcmp()`...

The `install_package()` function will be useful for other tests as well so it'd be better to put it somewhere in `lib/`. Preferably in a separate commit.

> > The `install_package()` function will be useful for other tests as well so it'd be better to put it somewhere in `lib/`. Preferably in a separate commit. > >...

`handle_screen()` has a few benefits over using `match_has_tag()` directly: - You don't have to make a long if/else code noodle to pick the right handler. If you have predefined handler...

Rebased on current master and updated verification run. This PR is ready for merge.