[LTP] [PATCH v1 1/4] lib/tst_kvercmp: Remove old distnames

[Jan Stancek]

On Tue, Dec 13, 2022 at 3:31 AM xuyang2018.jy@fujitsu.com
<xuyang2018.jy@fujitsu.com> wrote:
>
> Hi Jan, Petr
>
> > On Mon, Dec 12, 2022 at 4:30 PM Petr Vorel <pvorel@suse.cz> wrote:
> >>
> >> Hi,
> >>
> >>> Our travis-ci has dropped or never used them, they are useless.
> >>> So delete.
> >> They were here before CI. They were meant for using by tests
> >> (no CI related at all).
> >
> > Right, it's not just CI. RHEL5 is EOL, presumably Oracle as well, so
> > those likely won't be missed.
> > RHEL6 has 2 more years - Isn't this removal going to break tst_kvercmp2()?
>
> I have removed tst_kvercmp2 usage for RHEL6 in
> testcases/kernel/syscalls/inotify/inotify04.c and
> ../kernel/tracing/dynamic_debug/dynamic_debug01.sh, so it should not
> break tst_kvercmp2.

OK, you convinced me it's OK to drop it.

>
> Also, the current lastest ltp can't ensure that can be compile
> successfully on rhel6 because of lack of ci, so I think we don't need to
> still maintain the old version check.  If people want to use old kernel
> ie RHEL6, why not use the old ltp release?

They can, but they also likely want fixes to tests. I know there have been
some workarounds, such as using pre-configured LTP to workaround
autotools issue.

>
> ps: @Jan, if you want to keep RHEL6, I will keep it. But two more years,
> we still need to drop it.

You showed this isn't significant change for those few users that still might
exist (and EOL is coming), so:

Acked-by: Jan Stancek <jstancek@redhat.com>