mt76: allow VHT rate on 2.4GHz

[dengqf6]

Allow chips that support 11ac to use 256QAM on 2.4GHz
Ref: openwrt/openwrt#2522

[sgruszka]

Is this ok with IEEE 802.11 standard ?

[lukasz1992]

Not really https://www.cwnp.com/forums/posts?postNum=307375

[DavideFioravanti]

@dengqf6 To make it working with mt7615e I had to add this patch in /mt7615/init.c

@@ -457,6 +457,10 @@
 
 	mt7615_init_wiphy(hw);
 	dev->mphy.sband_2g.sband.ht_cap.cap |= IEEE80211_HT_CAP_LDPC_CODING;
+	dev->mphy.sband_2g.sband.vht_cap.cap |=
+			IEEE80211_VHT_CAP_MAX_MPDU_LENGTH_11454 |
+			IEEE80211_VHT_CAP_MAX_A_MPDU_LENGTH_EXPONENT_MASK |
+			IEEE80211_VHT_CAP_SUPP_CHAN_WIDTH_MASK | IEEE80211_VHT_CAP_SHORT_GI_160 | IEEE80211_VHT_CAP_SUPP_CHAN_WIDTH_160_80PLUS80MHZ;
 	dev->mphy.sband_5g.sband.ht_cap.cap |= IEEE80211_HT_CAP_LDPC_CODING;
 	dev->mphy.sband_5g.sband.vht_cap.cap |=
 			IEEE80211_VHT_CAP_MAX_MPDU_LENGTH_11454 |

It's really dirt and probably it should be done elsewhere, but without this patch hostapd wouldn't bring up the AP

[leonardcoutinho]

this can work in a mt7620a wifi chip?

[lukasz1992]

this can work in a mt7620a wifi chip?

definitely not

[Djfe]

Not really https://www.cwnp.com/forums/posts?postNum=307375

@lukasz1992 actually what about these days?
2.4GHz on 802.11ax goes up to 1024QAM.
Would a WiFi6 client (some Android device or Intel ax200) be able to connect to an older chip from the ac-age that does 802.11n with 256QAM on 2.4GHz?
I know that this isn't necessary. I'm just curious.
Or is it then called something else? (not VHT rate)

Edit:
also I feel like this PR can surely be closed since the other one in OpenWRT itself was closed due to lack of response by committers

[dhewg]

While this isn't covered by the specs, in practice this is used by quite some routers when using the vendor firmware.
Android supports it too, which doesn't make this exactly a small potion of the market.

For the record, I'm using this successfully on ipq40xx (so ath10k), and I rebased the patches of the mentioned closed PR, see openwrt/openwrt#2522 (comment)

[raenye]

To resolve the patch conflict, caused by a new 7th argument (bool ht) of mt76_init_sband, change the 2nd block to this:

@@ -349,13 +349,13 @@ mt76_init_sband(struct mt76_phy *phy, struct mt76_sband *msband,

 static int
 mt76_init_sband_2g(struct mt76_phy *phy, struct ieee80211_rate *rates,
-                  int n_rates)
+                  int n_rates, bool vht)
 {
        phy->hw->wiphy->bands[NL80211_BAND_2GHZ] = &phy->sband_2g.sband;

        return mt76_init_sband(phy, &phy->sband_2g, mt76_channels_2ghz,
                               ARRAY_SIZE(mt76_channels_2ghz), rates,
-                              n_rates, true, false);
+                              n_rates, true, vht);
 }

 static int

[V10lator]

Without this I get VHT on RX but not on TX. Giving speeds like this:

RX: 96.3 Mbit/s, 20 MHz, VHT-MCS 9, VHT-NSS 1, Short GI
TX: 65.0 Mbit/s, 20 MHz, MCS 7

So I think this should be implemented ASAP to give reasonable TX speeds.

[V10lator]

@dengqf6 To make it working with mt7615e I had to add this patch in /mt7615/init.c

New patch: V10lator@21e62be