> On 19. Jan 2023, at 12:35, Johannes Berg <johannes@xxxxxxxxxxxxxxxx> wrote:
>
> On Thu, 2023-01-19 at 12:31 +0100, Jonas Jelonek wrote:
>>> On 12. Jan 2023, at 11:26, Johannes Berg <johannes@xxxxxxxxxxxxxxxx>
>>> wrote:
>>>
>>>> +void ieee80211_rate_get_rate_info(const struct ieee80211_tx_rate *rate,
>>>> + struct wiphy *wiphy, u8 band,
>>>> + struct rate_info *rate_info)
>>>> +{
>>>> + memset(rate_info, 0, sizeof(struct rate_info));
>>>> +
>>>> + if (rate->flags & IEEE80211_TX_RC_MCS) { /* 802.11n */
>>>> + rate_info->flags |= RATE_INFO_FLAGS_MCS;
>>>> + rate_info->mcs = rate->idx;
>>>> + } else if (rate->flags & IEEE80211_TX_RC_VHT_MCS) { /* 802.11ac */
>>>> + rate_info->flags |= RATE_INFO_FLAGS_VHT_MCS;
>>>> + rate_info->mcs = ieee80211_rate_get_vht_mcs(rate);
>>>> + rate_info->nss = ieee80211_rate_get_vht_nss(rate);
>>>> + } else { /* 802.11a/b/g */
>>>
>>> what about HE/EHT?
>>
>> ieee80211_tx_rate uses an s8 for rate/MCS index, so only up to VHT
>> rates fit in there.
>> For rates above VHT, rate_info is needed, thus are are no HE/EHT rates
>> occuring in
>> ieee80211_tx_rate. Same applies to your comment on the hwsim
>> conversion.
>
> I guess I should've read the commit message more closely ;-)
>
> But please add kernel-doc to the function; both in general it'd be good
> to have, and in particular explaining that this is more for older
> drivers I guess?
I will add that in my next RFC version to make sure this is clear.
Jonas
