[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <3e9e4f3a-8968-4009-9e93-fe0bc8121392@kernel.org>
Date: Wed, 3 Jan 2024 09:36:06 -0700
From: David Ahern <dsahern@...nel.org>
To: nicolas.dichtel@...nd.com, "David S . Miller" <davem@...emloft.net>,
Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
Eric Dumazet <edumazet@...gle.com>, Phil Sutter <phil@....cc>
Cc: netdev@...r.kernel.org
Subject: Re: [PATCH net v2 2/2] selftests: rtnetlink: check enslaving iface in
a bond
On 1/3/24 9:17 AM, Nicolas Dichtel wrote:
> Le 03/01/2024 à 16:42, David Ahern a écrit :
>> On 1/3/24 2:48 AM, Nicolas Dichtel wrote:
>>> @@ -1239,6 +1240,44 @@ kci_test_address_proto()
>>> return $ret
>>> }
>>>
>>> +kci_test_enslave_bonding()
>>> +{
>>> + local testns="testns"
>>> + local bond="bond123"
>>> + local dummy="dummy123"
>>> + local ret=0
>>> +
>>> + run_cmd ip netns add "$testns"
>>> + if [ $? -ne 0 ]; then
>>> + end_test "SKIP bonding tests: cannot add net namespace $testns"
>>> + return $ksft_skip
>>> + fi
>>> +
>>> + # test native tunnel
>>> + run_cmd ip -netns $testns link add dev $bond type bond mode balance-rr
>>> + run_cmd ip -netns $testns link add dev $dummy type dummy
>>> + run_cmd ip -netns $testns link set dev $dummy up
>>> + run_cmd ip -netns $testns link set dev $dummy master $bond down
>>> + if [ $ret -ne 0 ]; then
>>> + end_test "FAIL: enslave an up interface in a bonding"
>>
>> interface is up, being put as a port on a bond and taken down at the
>> same time. That does not match this error message.
> The idea was "the command used to enslave an up interface fails".
> What about: "FAIL: set down and enslave an up interface in a bonding"
The 'set down' is part of the adding to the bond command. how about:
FAIL: Initially up interface added to a bond and set down.
Powered by blists - more mailing lists