[LTP] Refactoring doio testing suite

[Andrea Cervesato]

Hi all,

I'm trying to refactor the doio testing suite inside 
testcases/kernel/fs/doio.
This testing suite is pretty messy and difficult to read/maintain and 
for this
reason I thought to start the refactoring process. It also has the 
biggest amount
of code using old LTP libraries.

 From the point of view of the algorithm, everything is pretty simple 
and I managed
to go through all those tests arguments/headers etc, but the tlibio 
library is so messy
that I'm wondering if this task is worth the price.

For this reason I'm asking you how much we need this testing suite, 
considering that
fs is one of the biggest suites and it has multiple tests doing more or 
less everything,
from syscalls unit testing, to functional testing on async I/O, etc.

In short: according with your kernel maintenance experience, should we 
remove the
doio testing suite? I guess this is related also to how many bugs have 
been discovered
thanks to it.


Best regards,
Andrea Cervesato