[PATCH v3 8/8] test: rng: Add a UT testcase for the rng command

[Simon Glass]

Hi Sughosh,

On Sun, 13 Mar 2022 at 05:08, Sughosh Ganu <sughosh.ganu at linaro.org> wrote:
>
> hi Simon,
>
> On Sat, 12 Mar 2022 at 08:14, Simon Glass <sjg at chromium.org> wrote:
> >
> > Hi Sughosh,
> >
> > On Tue, 8 Mar 2022 at 23:18, Sughosh Ganu <sughosh.ganu at linaro.org> wrote:
> > >
> > > hi Simon,
> > >
> > > On Wed, 9 Mar 2022 at 08:06, Simon Glass <sjg at chromium.org> wrote:
> > > >
> > > > Hi Sughosh,
> > > >
> > > > On Fri, 4 Mar 2022 at 06:35, Sughosh Ganu <sughosh.ganu at linaro.org> wrote:
> > > > >
> > > > > The 'rng' command dumps a number of random bytes on the console. Add a
> > > > > set of tests for the 'rng' command. The test function performs basic
> > > > > sanity testing of the command.
> > > > >
> > > > > Since a unit test is being added for the command, enable it by default
> > > > > in the sandbox platforms.
> > > > >
> > > > > Signed-off-by: Sughosh Ganu <sughosh.ganu at linaro.org>
> > > > > ---
> > > > >
> > > > > Changes since V2: None
> > > > >
> > > > >  cmd/Kconfig   |  1 +
> > > > >  test/dm/rng.c | 29 +++++++++++++++++++++++++++++
> > > > >  2 files changed, 30 insertions(+)
> > > >
> > > > Reviewed-by: Simon Glass <sjg at chromium.org>
> > > >
> > > > >
> > > > > diff --git a/cmd/Kconfig b/cmd/Kconfig
> > > > > index 5e25e45fd2..47f1e23ef0 100644
> > > > > --- a/cmd/Kconfig
> > > > > +++ b/cmd/Kconfig
> > > > > @@ -1810,6 +1810,7 @@ config CMD_GETTIME
> > > > >  config CMD_RNG
> > > > >         bool "rng command"
> > > > >         depends on DM_RNG
> > > > > +       default y if SANDBOX
> > > > >         select HEXDUMP
> > > > >         help
> > > > >           Print bytes from the hardware random number generator.
> > > > > diff --git a/test/dm/rng.c b/test/dm/rng.c
> > > > > index 5b34c93ed6..6d1f68848d 100644
> > > > > --- a/test/dm/rng.c
> > > > > +++ b/test/dm/rng.c
> > > > > @@ -25,3 +25,32 @@ static int dm_test_rng_read(struct unit_test_state *uts)
> > > > >         return 0;
> > > > >  }
> > > > >  DM_TEST(dm_test_rng_read, UT_TESTF_SCAN_PDATA | UT_TESTF_SCAN_FDT);
> > > > > +
> > > > > +/* Test the rng command */
> > > > > +static int dm_test_rng_cmd(struct unit_test_state *uts)
> > > > > +{
> > > > > +       struct udevice *dev;
> > > > > +
> > > > > +       ut_assertok(uclass_get_device(UCLASS_RNG, 0, &dev));
> > > > > +       ut_assertnonnull(dev);
> > > > > +
> > > > > +       ut_assertok(console_record_reset_enable());
> > > > > +
> > > > > +       run_command("rng", 0);
> > > > > +       ut_assert_nextlinen("00000000:");
> > > > > +       ut_assert_nextlinen("00000010:");
> > > > > +       ut_assert_nextlinen("00000020:");
> > > > > +       ut_assert_nextlinen("00000030:");
> > > >
> > > > This is good enough for testing, except that here you don't have any
> > > > actual data. How come?
> > >
> > > This being a test for a random number device, we cannot anticipate
> > > what data we will be getting. The sandbox RNG driver does not return a
> > > fixed set of values on each invocation of the command.
> >
> > How come? Can you fix that by setting the random seed at the start of the test?
>
> Well I can, but do we really need to do that. I think the current test
> that we have for the rng command and the uclass is fine.

It is OK, but not great, since if there is a bug in the data output or
in the collection of the random data, then you will not find it with
this test.

Regards,
Simon