[PATCH v2 6/7] dts: revamp runtime internals

Luca Vizzarro Luca.Vizzarro at arm.com
Mon Feb 17 11:26:21 CET 2025


On 14/02/2025 18:54, Nicholas Pratte wrote:
> This definitely cleans things up a lot!
> 
> Reviewed-by: Nicholas Pratte <npratte at iol.unh.edu>
> 
> <snip>
>> @@ -217,9 +218,7 @@ def get_supported_capabilities(
>>           )
>>           if cls.capabilities_to_check:
>>               capabilities_to_check_map = cls._get_decorated_capabilities_map()
>> -            with TestPmdShell(
>> -                sut_node, privileged=True, disable_device_start=True
>> -            ) as testpmd_shell:
>> +            with TestPmdShell() as testpmd_shell:
> 
> This is probably a needless concern, but should this change
> technically be in the last patch in order for it to stand on its own?
> Not sure if that really matters or not, but figured I'd bring it up.

No, not really, because the API change happens here. I removed 
privileged because it's redundant, and disabled_device_start as it 
seemed problematic for some things. We should handle commands stop the 
ports with @require_stopped_ports, so that we have a consistent 
execution. This is how I discovered the problem with the Mellanox and 
the MTU setting.


More information about the dev mailing list