[PATCH i-g-t 5/5] tests/dumb_buffer: Add subtests for mixing dumb buffers with prime

[Zack Rusin]

On Fri, Jun 28, 2024 at 12:03 PM Kamil Konieczny
<kamil.konieczny at linux.intel.com> wrote:
>
> Hi Zack,
> On 2024-06-27 at 15:56:15 -0400, Zack Rusin wrote:
> > New versions of KWin are extensively using dumb buffers with prime. Add
> > some basic tests to verify that mixing the dumb buffer interface
> > with prime ends up with a well formed buffer.
> >
> > Signed-off-by: Zack Rusin <zack.rusin at broadcom.com>
> > ---
> >  tests/dumb_buffer.c | 96 +++++++++++++++++++++++++++++++++++++++++++++
> >  1 file changed, 96 insertions(+)
> >
>
> This seems unrelated to your series, please send it in separate
> one. Few more nits below.

Sounds good. It's generally just improvements to prime/dmabuf testing
framework in igt, but I'm happy to send that as a separate patch.

> > diff --git a/tests/dumb_buffer.c b/tests/dumb_buffer.c
> > index 8e243459c..0651f5d83 100644
> > --- a/tests/dumb_buffer.c
> > +++ b/tests/dumb_buffer.c
> > @@ -69,6 +69,20 @@
> >   * SUBTEST: map-uaf
> >   *
> >   * SUBTEST: map-valid
> > + *
> > + * SUBTEST: dmabuf-read
> > + * Description: Make a dumb buffer, map it using the dumb buffer interface,
> > + *              export it via prime, mmap the prime buffer and check
> > + *              that its contents matches what we wrote to the dumb
> > + *              buffer.
> > + * Functionality: dumb_buffers, prime, dmabuf, mmap
> ----- ^
> Do you need it here?

We don't. I just looked at other files e.g. kms_flip.c and kms_prime.c
which seem to list "Functionality" for the added tests so I was trying
to follow what I assumed best practice was. I'm happy to remove any
part of the docs for it.

z