Does _need_to_be_root deserve to live?

Does _need_to_be_root deserve to live?

[Theodore Ts'o]

There appear to be a number of tests that need to mount and unmount
scratch or su to $qa_user that don't call the helper _need_to_be_root.

Also, sicne the check script calls _check_filesystems between every
single test, which requires unmounting and remounting the test and/or
scratch file systems, does the _need_to_be_root check make any sense
at all?

And if doesn't make any sense, is it worth it to submitting a patch
that just removes all mention of _need_to_be_root from the tests and
common/rc?

Cheers,

						- Ted

Re: Does _need_to_be_root deserve to live?

[Eryu Guan]

On Wed, Feb 10, 2016 at 10:57:46PM -0500, Theodore Ts'o wrote:
> There appear to be a number of tests that need to mount and unmount
> scratch or su to $qa_user that don't call the helper _need_to_be_root.
> 
> Also, sicne the check script calls _check_filesystems between every
> single test, which requires unmounting and remounting the test and/or
> scratch file systems, does the _need_to_be_root check make any sense
> at all?
> 
> And if doesn't make any sense, is it worth it to submitting a patch
> that just removes all mention of _need_to_be_root from the tests and
> common/rc?

I think it's safe to remove all _need_to_be_root calls, because there's
already a user-id check in "check"

346 if [ `id -u` -ne 0 ]
347 then
348     echo "check: QA must be run as root"
349     exit 1
350 fi

Thanks,
Eryu