Skip to contentSkip to conversation

ath11k: HE mesh support

RFC PATCH Superseded Sven Eckelmann 2 participants 4 messages Aug 13, 2019 – Sep 19, 2019 Qualcomm Atheros ATH11K wireless driver

v6 of 7 Superseded Latest is v7 →

lore
Share & tools
lore
#0 lore raw

v6 cover ath11k: HE mesh support

From: Sven Eckelmann <[email protected]>

Hi,

Some features of 802.11ax without central organizing (AP) STA can also be used in mesh mode. The main goal was to get HE mesh working with ath11k. But the support for HE was completely removed from ath11k [1]. So these RFC patches are just stuff which doesn't apply anymore because of this and are just dropped here to have some (more or less) clean dump in case HE is added again at some point.

Kind regards,
Sven

[1] https://git.kernel.org/pub/scm/linux/kernel/git/kvalo/ath.git/commit/?id=50ca7786c26fba9db211ed89843e0da47b0cfbc0

* v6

- remove upstream merged mac80211 patch
- send only the ath11k relevant patches
- drop (internal DEP-3-like) "Forwarded:" tag
- switch from PATCH to RFC due to dropped HE support in ath11k

* v5

- rebased patches
- moved ath11k meshpoint HE capabilities filter to avoid lines over 80 characters
- add patch to disable HE SU phy capabilities to work around problems

    with WLAN.HK.2.1.0.1-00629-QCAHKSWPL_SILICONZ-1 firmware

* v4

- switch order of the mac80211 HE mesh support and mac80211_hwsim support
- drop ie_len variable from ieee80211_ie_len_he_cap and

    ieee80211_ie_build_he_oper
  - switch to function ieee80211_get_he_iftype_cap instead of implementing new
    function ieee80211_get_he_mesh_cap

- allow ieee80211_ie_len_he_cap to calculate length for non-mesh HE cap
- dropped he_support workaround in ath11k because now all phy_modes are HE phy modes

* v3

- force ath11k PHY mode for meshpoint vif to HE mode to avoid hang of firmware when HE (or VHT on 2.4GHz) device tries to connect

* v2:

- add of ath11k patch

* v1:

- initial RFC

Sven Eckelmann (2):
ath11k: register HE mesh capabilities
ath11k: Disable SU support in meshpoint mode

 drivers/net/wireless/ath/ath11k/mac.c | 84 +++++++++++++++++++++++++++
 1 file changed, 84 insertions(+)
Signature · 7 lines
2.20.1


_______________________________________________
ath11k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath11k
#1 lore raw

RFC PATCH v6 1/2 ath11k: register HE mesh capabilities

From: Sven Eckelmann <[email protected]>

The capabilities for the HE mesh are generated from the capabilities reported by the fw. But the firmware only reports the overall capabilities and not the one which are specific for mesh. Some of them (TWT, MU UL/DL, TB PPDU, ...) require an infrastructure setup with a main STA (AP) controlling the operations. This is not the case for mesh and thus these capabilities are removed from the list of capabilities.

---
 drivers/net/wireless/ath/ath11k/mac.c | 75 +++++++++++++++++++++++++++
 1 file changed, 75 insertions(+)
drivers/net/wireless/ath/ath11k/mac.c +75 -0
index 1ade831070b4..9abe4a25a1b6 100644
@@ -3264,6 +3264,77 @@ static void ath11k_gen_ppe_thresh(struct ath11k_ppe_threshold *fw_ppet,
 	}
 }
 
+static void
+ath11k_mac_filter_he_cap_mesh(struct ieee80211_he_cap_elem *he_cap_elem)
+{
+	u8 m;
+
+	m = IEEE80211_HE_MAC_CAP0_TWT_RES |
+	    IEEE80211_HE_MAC_CAP0_TWT_REQ;
+	he_cap_elem->mac_cap_info[0] &= ~m;
+
+	m = IEEE80211_HE_MAC_CAP2_TRS |
+	    IEEE80211_HE_MAC_CAP2_BCAST_TWT |
+	    IEEE80211_HE_MAC_CAP2_MU_CASCADING;
+	he_cap_elem->mac_cap_info[2] &= ~m;
+
+	m = IEEE80211_HE_MAC_CAP3_FLEX_TWT_SCHED |
+	    IEEE80211_HE_MAC_CAP2_BCAST_TWT |
+	    IEEE80211_HE_MAC_CAP2_MU_CASCADING;
+	he_cap_elem->mac_cap_info[3] &= ~m;
+
+	m = IEEE80211_HE_MAC_CAP4_BSRP_BQRP_A_MPDU_AGG |
+	    IEEE80211_HE_MAC_CAP4_BQR;
+	he_cap_elem->mac_cap_info[4] &= ~m;
+
+	m = IEEE80211_HE_MAC_CAP5_SUBCHAN_SELECVITE_TRANSMISSION |
+	    IEEE80211_HE_MAC_CAP5_UL_2x996_TONE_RU |
+	    IEEE80211_HE_MAC_CAP5_PUNCTURED_SOUNDING |
+	    IEEE80211_HE_MAC_CAP5_HT_VHT_TRIG_FRAME_RX;
+	he_cap_elem->mac_cap_info[5] &= ~m;
+
+	m = IEEE80211_HE_PHY_CAP2_UL_MU_FULL_MU_MIMO |
+	    IEEE80211_HE_PHY_CAP2_UL_MU_PARTIAL_MU_MIMO;
+	he_cap_elem->phy_cap_info[2] &= ~m;
+
+	m = IEEE80211_HE_PHY_CAP3_RX_HE_MU_PPDU_FROM_NON_AP_STA |
+	    IEEE80211_HE_PHY_CAP3_DCM_MAX_CONST_TX_MASK |
+	    IEEE80211_HE_PHY_CAP3_DCM_MAX_CONST_RX_MASK;
+	he_cap_elem->phy_cap_info[3] &= ~m;
+
+	m = IEEE80211_HE_PHY_CAP4_MU_BEAMFORMER;
+	he_cap_elem->phy_cap_info[4] &= ~m;
+
+	m = IEEE80211_HE_PHY_CAP5_NG16_MU_FEEDBACK;
+	he_cap_elem->phy_cap_info[5] &= ~m;
+
+	m = IEEE80211_HE_PHY_CAP6_CODEBOOK_SIZE_75_MU |
+	    IEEE80211_HE_PHY_CAP6_TRIG_MU_BEAMFORMER_FB |
+	    IEEE80211_HE_PHY_CAP6_TRIG_CQI_FB |
+	    IEEE80211_HE_PHY_CAP6_PARTIAL_BANDWIDTH_DL_MUMIMO;
+	he_cap_elem->phy_cap_info[6] &= ~m;
+
+	m = IEEE80211_HE_PHY_CAP7_SRP_BASED_SR |
+	    IEEE80211_HE_PHY_CAP7_POWER_BOOST_FACTOR_AR |
+	    IEEE80211_HE_PHY_CAP7_STBC_TX_ABOVE_80MHZ |
+	    IEEE80211_HE_PHY_CAP7_STBC_RX_ABOVE_80MHZ;
+	he_cap_elem->phy_cap_info[7] &= ~m;
+
+	m = IEEE80211_HE_PHY_CAP8_HE_ER_SU_PPDU_4XLTF_AND_08_US_GI |
+	    IEEE80211_HE_PHY_CAP8_20MHZ_IN_40MHZ_HE_PPDU_IN_2G |
+	    IEEE80211_HE_PHY_CAP8_20MHZ_IN_160MHZ_HE_PPDU |
+	    IEEE80211_HE_PHY_CAP8_80MHZ_IN_160MHZ_HE_PPDU;
+	he_cap_elem->phy_cap_info[8] &= ~m;
+
+	m = IEEE80211_HE_PHY_CAP9_LONGER_THAN_16_SIGB_OFDM_SYM |
+	    IEEE80211_HE_PHY_CAP9_NON_TRIGGERED_CQI_FEEDBACK |
+	    IEEE80211_HE_PHY_CAP9_RX_1024_QAM_LESS_THAN_242_TONE_RU |
+	    IEEE80211_HE_PHY_CAP9_TX_1024_QAM_LESS_THAN_242_TONE_RU |
+	    IEEE80211_HE_PHY_CAP9_RX_FULL_BW_SU_USING_MU_WITH_COMP_SIGB |
+	    IEEE80211_HE_PHY_CAP9_RX_FULL_BW_SU_USING_MU_WITH_NON_COMP_SIGB;
+	he_cap_elem->phy_cap_info[9] &= ~m;
+}
+
 static int ath11k_mac_copy_he_cap(struct ath11k *ar,
 				  struct ath11k_pdev_cap *cap,
 				  struct ieee80211_sband_iftype_data *data,
@@ -3280,6 +3351,7 @@ static int ath11k_mac_copy_he_cap(struct ath11k *ar,
 		switch (i) {
 		case NL80211_IFTYPE_STATION:
 		case NL80211_IFTYPE_AP:
+		case NL80211_IFTYPE_MESH_POINT:
 			break;
 
 		default:
@@ -3320,6 +3392,9 @@ static int ath11k_mac_copy_he_cap(struct ath11k *ar,
 			he_cap_elem->phy_cap_info[9] |=
 				IEEE80211_HE_PHY_CAP9_TX_1024_QAM_LESS_THAN_242_TONE_RU;
 			break;
+		case NL80211_IFTYPE_MESH_POINT:
+			ath11k_mac_filter_he_cap_mesh(he_cap_elem);
+			break;
 		}
 
 		he_cap->he_mcs_nss_supp.rx_mcs_80 =
Signature · 7 lines
2.20.1


_______________________________________________
ath11k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath11k
#2 lore raw

RFC PATCH v6 2/2 ath11k: Disable SU support in meshpoint mode

In reply to Sven Eckelmann

From: Sven Eckelmann <[email protected]>

The firmware WLAN.HK.2.1.0.1-00629-QCAHKSWPL_SILICONZ-1 often stops sending data on 5GHz when SU support is advertised. A simple throughput benchmark will end up at only a couple hundred kilobytes per second (when finishing at all).

Disabling SU related flags works around this problem and allowed a maximum throughput of ~450 MBit/s in a simple test setup with an HK01 over air.

---
 drivers/net/wireless/ath/ath11k/mac.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)
drivers/net/wireless/ath/ath11k/mac.c +12 -3
index 9abe4a25a1b6..66be6c8faad1 100644
@@ -3299,16 +3299,24 @@ ath11k_mac_filter_he_cap_mesh(struct ieee80211_he_cap_elem *he_cap_elem)
 
 	m = IEEE80211_HE_PHY_CAP3_RX_HE_MU_PPDU_FROM_NON_AP_STA |
 	    IEEE80211_HE_PHY_CAP3_DCM_MAX_CONST_TX_MASK |
-	    IEEE80211_HE_PHY_CAP3_DCM_MAX_CONST_RX_MASK;
+	    IEEE80211_HE_PHY_CAP3_DCM_MAX_CONST_RX_MASK |
+	    IEEE80211_HE_PHY_CAP3_SU_BEAMFORMER;
 	he_cap_elem->phy_cap_info[3] &= ~m;
 
-	m = IEEE80211_HE_PHY_CAP4_MU_BEAMFORMER;
+	m = IEEE80211_HE_PHY_CAP4_SU_BEAMFORMEE |
+	    IEEE80211_HE_PHY_CAP4_MU_BEAMFORMER |
+	    IEEE80211_HE_PHY_CAP4_BEAMFORMEE_MAX_STS_UNDER_80MHZ_MASK |
+	    IEEE80211_HE_PHY_CAP4_BEAMFORMEE_MAX_STS_ABOVE_80MHZ_MASK;
 	he_cap_elem->phy_cap_info[4] &= ~m;
 
-	m = IEEE80211_HE_PHY_CAP5_NG16_MU_FEEDBACK;
+	m = IEEE80211_HE_PHY_CAP5_BEAMFORMEE_NUM_SND_DIM_UNDER_80MHZ_MASK |
+	    IEEE80211_HE_PHY_CAP5_BEAMFORMEE_NUM_SND_DIM_ABOVE_80MHZ_MASK |
+	    IEEE80211_HE_PHY_CAP5_NG16_SU_FEEDBACK |
+	    IEEE80211_HE_PHY_CAP5_NG16_MU_FEEDBACK;
 	he_cap_elem->phy_cap_info[5] &= ~m;
 
 	m = IEEE80211_HE_PHY_CAP6_CODEBOOK_SIZE_75_MU |
+	    IEEE80211_HE_PHY_CAP6_TRIG_SU_BEAMFORMER_FB |
 	    IEEE80211_HE_PHY_CAP6_TRIG_MU_BEAMFORMER_FB |
 	    IEEE80211_HE_PHY_CAP6_TRIG_CQI_FB |
 	    IEEE80211_HE_PHY_CAP6_PARTIAL_BANDWIDTH_DL_MUMIMO;
@@ -3316,6 +3324,7 @@ ath11k_mac_filter_he_cap_mesh(struct ieee80211_he_cap_elem *he_cap_elem)
 
 	m = IEEE80211_HE_PHY_CAP7_SRP_BASED_SR |
 	    IEEE80211_HE_PHY_CAP7_POWER_BOOST_FACTOR_AR |
+	    IEEE80211_HE_PHY_CAP7_HE_SU_MU_PPDU_4XLTF_AND_08_US_GI |
 	    IEEE80211_HE_PHY_CAP7_STBC_TX_ABOVE_80MHZ |
 	    IEEE80211_HE_PHY_CAP7_STBC_RX_ABOVE_80MHZ;
 	he_cap_elem->phy_cap_info[7] &= ~m;
Signature · 7 lines
2.20.1


_______________________________________________
ath11k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath11k
#3 lore raw

Re: [RFC PATCH v6 2/2] ath11k: Disable SU support in meshpoint mode

Quoted 11 lines · On Tuesday, 13 August 2019 11:27:26 CEST Sven Eckelmann wrote:

From: Sven Eckelmann <[email protected]>

The firmware WLAN.HK.2.1.0.1-00629-QCAHKSWPL_SILICONZ-1 often stops sending data on 5GHz when SU support is advertised. A simple throughput benchmark will end up at only a couple hundred kilobytes per second (when finishing at all).

Disabling SU related flags works around this problem and allowed a maximum throughput of ~450 MBit/s in a simple test setup with an HK01 over air.

I've just checked with WLAN.HK.2.0.0.1-00043-QCAHKSWPL_SILICONZ-1 (which is a newer version than WLAN.HK.2.1.0.1-00629-QCAHKSWPL_SILICONZ-1) and wasn't able anymore to reproduce the SU related problems on 5GHz.

Kind regards,
Sven

_______________________________________________
ath11k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath11k