[LTP] [PATCH v2] kernel/fs/fsnotify-stress: fsnotify stress test

[Petr Vorel]

> This is a stress tests that exercises fanotify and inotify interfaces
> while IO going on. It ignores some failures of syscalls to let the
> stress go on. If the kernel does not panic or hang after a certain
> period of time of testing, test pass.

> Signed-off-by: Murphy Zhou <jencce.kernel@gmail.com>
> ---
> v2:
> 	Merge all into one C programme.
> 	Changed dirname of the testcase.
> 	Run fs_racer.sh from this.
Also tests in testcases/kernel/fs/racer/ are very old scripts with questionable
code quality. I would not use them with new tests.  Not sure whether these tests
are useful at all, but if yes they please take the idea and implement it in new
C API, so that they can be called by forked thread.

Why C? Running shell test from C is probably not acceptable. We have enough
problems with shell tests which use C binaries. Doing it vice versa can bring
even more problems.

fs_racer.sh runs many times fs_racer_file_create.sh, which mainly does:
dd if=/dev/zero of=$DIR/$file bs=1k count=$SIZE

=> nothing hard to implement in C.

Kind regards,
Petr