Re: [PATCH net-next 1/2] net: ptp: introduce .supported_extts_flags to ptp_clock_info

[Jacob Keller]

On 4/11/2025 6:20 PM, Jakub Kicinski wrote:
> Sorry for the late nit but the conversion is pretty inconsistent..
> 
> On Tue, 08 Apr 2025 13:55:14 -0700 Jacob Keller wrote:

8<--

>> @@ -1378,6 +1368,8 @@ void igb_ptp_init(struct igb_adapter *adapter)
>>  		adapter->ptp_caps.n_ext_ts = IGB_N_EXTTS;
>>  		adapter->ptp_caps.n_per_out = IGB_N_PEROUT;
>>  		adapter->ptp_caps.n_pins = IGB_N_SDP;
>> +		adapter->ptp_caps.supported_extts_flags = PTP_EXTTS_EDGES |
>> +							  PTP_STRICT_FLAGS;
> 
> sometimes you both drop the enabled and combine the edges 
> 
>>  		adapter->ptp_caps.pps = 1;
>>  		adapter->ptp_caps.pin_config = adapter->sdp_config;
>>  		adapter->ptp_caps.adjfine = igb_ptp_adjfine_82580;
> 
> No preference which version you pick but shouldn't we go with one?
> Or is this on purpose to show we have no preference?

It wasn't on purpose. Honestly not sure how they ended up different. I
think the right thing is:

a) drop the ENABLE since we always accept that regardless of whether its
in supported_extts_flags

b) don't use PTP_EXTTS_EDGES in the supported flags, since its easier to
see which specific flags are supported if you aren't having to remember
that the macro means multiple flags.

I'll update in a v2 to be consistent.

Thanks,
Jake