> * The number of peers that HW timestamping can be enabled for concurrently
> * is indicated by %NL80211_ATTR_MAX_HW_TIMESTAMP_PEERS.
> *
> + * @NL80211_CMD_UPDATE_HE_MUEDCA_PARAMS: Updated MU-EDCA parameters from
> + * driver. This event is used to update dynamic MU-EDCA parameters in
> + * management frames, coming from driver and now need to be reflected in
> + * management frames.
Not sure I'm happy with this description - it has all the right
information but feels a bit odd? And the first sentence isn't even a
valid sentence. Please rephrase.
> + * @NL80211_ATTR_HE_MUEDCA_PARAMS: MU-EDCA AC parameters for the
> + * %NL80211_CMD_UPDATE_HE_MUEDCA_PARAMS command in format described in
> + * P802.11ax_D4.0 section 9.4.2.245
"in the format described" or maybe "as described"?
> diff --git a/net/wireless/nl80211.c b/net/wireless/nl80211.c
> index e9e939706630..af36440f2a51 100644
> --- a/net/wireless/nl80211.c
> +++ b/net/wireless/nl80211.c
> @@ -816,6 +816,8 @@ static const struct nla_policy nl80211_policy[NUM_NL80211_ATTR] = {
> [NL80211_ATTR_HW_TIMESTAMP_ENABLED] = { .type = NLA_FLAG },
> [NL80211_ATTR_EMA_RNR_ELEMS] = { .type = NLA_NESTED },
> [NL80211_ATTR_DYN_MUEDCA_ENABLE] = { .type = NLA_FLAG },
> + [NL80211_ATTR_HE_MUEDCA_PARAMS] =
> + NLA_POLICY_EXACT_LEN(sizeof(struct ieee80211_mu_edca_param_set)),
This is output only, so arguably should be NLA_REJECT, or just not
listed?
> +EXPORT_SYMBOL(cfg80211_update_muedca_params_event);
And again here (adding Jeff and Jouni), we have an exported cfg80211
function that, as far as I can tell, you have no plans whatsoever to use
in an upstream driver. Yes it's only 130 lines of code or so, but still,
someone has to maintain it ...
So please resend this only when you have a user. Until then you can
trivially use a vendor event for this and I don't have to care, and I
don't even have to review it and address all the stupid little issues
like missing documentation that you'd have found if you had any internal
review worth calling that.
johannes
