Re: [PATCH] xarray: port tests to kunit

[Tamir Duberstein <tamird@gmail.com>]

On Wed, Jan 29, 2025 at 5:26 PM Liam R. Howlett <Liam.Howlett@oracle.com> wrote:
>
> * Tamir Duberstein <tamird@gmail.com> [250129 16:29]:
> > On Wed, Jan 29, 2025 at 4:26 PM Liam R. Howlett <Liam.Howlett@oracle.com> wrote:
> > >
> > > * Sidhartha Kumar <sidhartha.kumar@oracle.com> [250129 16:02]:
> > > > + Liam, Matthew
> > >
> > > + linux-mm
> > >
> > > Thank you Sid.
> > >
> > > >
> > > > Hello,
> > > >
> > > > I believe this patch needs to be reverted for now as it breaks the
> > > > user-space build of /tools/testing/radix-tree with:
> > > >
> > > > In file included from xarray.c:11:
> > > > ../../../lib/test_xarray.c:9:10: fatal error: kunit/test.h: No such file
> > > > or directory
> > > >      9 | #include <kunit/test.h>
> > > >        |          ^~~~~~~~~~~~~~
> > > > compilation terminated.
> > > > make: *** [<builtin>: xarray.o] Error 1
> > > > make: *** Waiting for unfinished jobs....
> > > >
> > > > this then prevents the maple tree test suite from building.
> > >
> > > How are grammar corrections going to the right person (but not the
> > > mailing list) while an entire conversion to kunit is not [1]?
> >
> > Very simple: the tests are not properly included in MAINTAINERS. I
> > sent https://lore.kernel.org/all/20250129-xarray-test-maintainer-v1-1-482e31f30f47@gmail.com/
> > a few minutes ago for this reason.
>
> Fair enough, but from the patch:
>
> @@ -6,11 +6,10 @@
>   * Author: Matthew Wilcox <willy@infradead.org>
>   */
>
> -#include <linux/xarray.h>
> -#include <linux/module.h>
> +#include <kunit/test.h>
>
> ...
>
>
> -module_init(xarray_checks);
> -module_exit(xarray_exit);
>  MODULE_AUTHOR("Matthew Wilcox <willy@infradead.org>");
>  MODULE_DESCRIPTION("XArray API test module");
>  MODULE_LICENSE("GPL");
>
> I don't get why the huge list of Cc's didn't include the author who is
> in the git commit signers:
>  $ ./scripts/get_maintainer.pl --git lib/xarray.c
> Matthew Wilcox <willy@infradead.org> (supporter:XARRAY,commit_signer:1/3=33%,authored:1/3=33%,added_lines:19/52=37%,removed_lines:4/23=17%)
> Andrew Morton <akpm@linux-foundation.org> (supporter:LIBRARY CODE,commit_signer:3/3=100%)

I'm not sure what you're asking. I used `b4 prep --auto-to-cc`. It
doesn't know that test_xarray.c and xarray.c have the same maintainer.

> ...
>
> >
> > > Does the patch really need to drop the module testing too?
> > >
> > > What exactly is the point of converting one testing system to another
> > > besides disruption of actual work?  Who asked for this?  What is the
> > > point?
> >
> > All this is described in the commit message.
>
> The commit message says you like the output more and implies you like
> the command better.
>
> I've never used the kunit testing of xarray and have used the userspace
> testing instead, so I can't speak to the obscure invocation as both
> commands seem insanely long and obscure to me.
>
> >
> > > Is anyone doing work on the xarray running the kunit tests?
> >
> > I was doing work on xarray and I was running the kunit tests.
>
> ...
>
> You should look at the userspace testing (that this broke) as it has
> been really useful in certain scenarios.
>
> Thanks,
> Liam

For what it's worth the kunit invocation, while obscure, is
self-documenting. There's usage information that's reasonably
understandable embedded in the tool itself. I looked for the userspace
testing initially but failed to find
tools/testing/radix-tree/xarray.c. Even now, I'm not sure how I'm
meant to compile this.

Tamir