On 5/1/25 09:01, Ritesh Harjani (IBM) wrote:
"Nirjhar Roy (IBM)" <nirjhar.roy.lists@xxxxxxxxx> writes:
Some of the "status=<val>;exit" and "exit <val>" were not
replaced with _exit <val> and _fatal. Doing it now.
Indeed a nice cleanup. The changes in this patch looks good to me.
Please feel free to add:
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@xxxxxxxxx>
Thank you.
So I guess these couple of series was to cleanup exit routines from
common bash scripts. Do we plan to update the tests/ as well where
we call...
status=X
exit
...or updating tests/ is not needed since we didn't find any wrong usage of
"exit X" routines there?
Thank you for pointing this out. The exit command is used in 2 ways in
the tests:
1. "exit 1"
2. "status=0; exit"
1) works because we set the value of "status" to 1 (failure by default)
in _begin_fstest() - so even if "exit 1" is not correctly explicitly
setting the value of "status", it simply works. However, "exit <any
value != 1>" will not work (although I didn't find any place in tests
where exit has been used with any other value apart from 0 and 1).
2) This works since we are setting the value of "status" correcting
before "exit"ing.
But yes, we should ideally replace direct usage of exit with either
_exit or _fatal (depending on the exit value). I will add this to my
ToDo list and send a separate patch series with this and the README
change you have suggested below.
Either ways - I think we might need to update the README at some point
in time which carries this snip. You might need to add that there are
helper routines like _exit() and _fatal() perhaps for use in common
scripts.
<snip>
To force a non-zero exit status use:
status=1
exit
Note that:
exit 1
won't have the desired effect because of the way the exit trap
works.
I agree. I will send a separate patch with this and the exit call
replacement of the tests.
--NR
-ritesh
--
Nirjhar Roy
Linux Kernel Developer
IBM, Bangalore
