[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 4 Apr 2023 12:34:03 -0400
From: Jonathan Toppins <jtoppins@...hat.com>
To: Hangbin Liu <liuhangbin@...il.com>
Cc: netdev@...r.kernel.org, Jay Vosburgh <j.vosburgh@...il.com>,
"David S . Miller" <davem@...emloft.net>,
Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>,
Eric Dumazet <edumazet@...gle.com>, Liang Li <liali@...hat.com>
Subject: Re: [PATCH net 2/3] selftests: bonding: re-format bond option tests
On 4/3/23 22:28, Hangbin Liu wrote:
> On Fri, Mar 31, 2023 at 10:31:47AM +0800, Hangbin Liu wrote:
>>>> +++ b/tools/testing/selftests/drivers/net/bonding/bond_lib.sh
>>>
>>> I like this idea, we might want to separate network topology from library
>>> code however. That way a given test case can just include a predefined
>>
>> Would you like to help explain more clear? Separate network topology to where?
>
>
> Hi Jon, would you please help explain this part?
Thanks for the ping. It looks like several test cases build largely the
same virtual network topology and then execute the test case. I was
attempting to point out that it might be better to provide a standard
network topology and then each test case utilizes this standard topology
instead of each test case rolling its own. Also, with my comment about
separating out the topology from library code I was accounting for the
ability to support multiple topologies, fe:
bond_lib.sh
bond_topo_gateway.sh
bond_topo_2.sh
Then a given test case only includes/sources `bond_topo_gateway.sh`
which creates the virtual network.
-Jon
>
>>
>>> topology. A quick review of the test cases show a 2 node setup is the most
>>> common across all test cases.
>>
>> Liang suggested that with 2 clients we can test xmit_hash_policy. In
>> client_create() I only create 1 client for current testing. We can add more
>> clients in future.
>>
>> Thanks
>> Hangbin
>
Powered by blists - more mailing lists