MIME-Version: 1.0
References: <20230306111959.429680-1-po-hsu.lin@canonical.com> <20230306103316.3224383e@kernel.org>
In-Reply-To: <20230306103316.3224383e@kernel.org>
From:   Po-Hsu Lin <po-hsu.lin@canonical.com>
Date:   Tue, 7 Mar 2023 10:24:25 +0800
Message-ID: <CAMy_GT89khdMOCeSxqGx_hh+samb0BabbrQrY6nd9uA3xNS4Cg@mail.gmail.com>
Subject: Re: [PATCH] selftests: net: devlink_port_split.py: skip test if no
 suitable device available
To:     Jakub Kicinski <kuba@kernel.org>
Cc:     linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
        netdev@vger.kernel.org, idosch@mellanox.com,
        danieller@mellanox.com, petrm@mellanox.com, shuah@kernel.org,
        pabeni@redhat.com, edumazet@google.com, davem@davemloft.net
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Tue, Mar 7, 2023 at 2:33=E2=80=AFAM Jakub Kicinski <kuba@kernel.org> wro=
te:
>
> On Mon,  6 Mar 2023 19:19:59 +0800 Po-Hsu Lin wrote:
> > The `devlink -j dev show` command output may not contain the "flavour"
> > key, for example:
> >   $ devlink -j dev show
> >   {"dev":{"pci/0001:00:00.0":{},"pci/0002:00:00.0":{}}}
>
> It's not dev that's supposed to have the flavor, it's port.
>
>   devlink -j port show
Ah yes, it's using output from this command, thanks for catching this.

>
> Are you running with old kernel or old user space?
> Flavor is not an optional attribute.
This was from a s390x LPAR instance with Ubuntu 22.10 (5.19.0-37-generic)
iproute2-5.15.0

$ devlink -j port show
{"port":{"pci/0001:00:00.0/1":{"type":"eth","netdev":"ens301"},"pci/0001:00=
:00.0/2":{"type":"eth","netdev":"ens301d1"},"pci/0002:00:00.0/1":{"type":"e=
th","netdev":"ens317"},"pci/0002:00:00.0/2":{"type":"eth","netdev":"ens317d=
1"}}}

>
> > This will cause a KeyError exception. Fix this by checking the key
> > existence first.
> >
> > Also, if max lanes is 0 the port splitting won't be tested at all.
> > but the script will end normally and thus causing a false-negative
> > test result.
> >
> > Use a test_ran flag to determine if these tests were skipped and
> > return KSFT_SKIP accordingly.
> >
> > Link: https://bugs.launchpad.net/bugs/1937133
> > Fixes: f3348a82e727 ("selftests: net: Add port split test")
> > Signed-off-by: Po-Hsu Lin <po-hsu.lin@canonical.com>
>
> Could you factor out the existing skipping logic from main()
> (the code under "if not dev:") and add the test for flavors
> to the same function? It'll be a bit more code but cleaner
> result IMHO.
Sure, will do with V2.
Thanks!