openwrt/package/kernel/mac80211/patches/rt2x00/010-rt2x00-add-TX-LOFT-calibration-for-MT7620.patch

983 lines
33 KiB
Diff
Raw Normal View History

mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
From patchwork Sat Sep 17 20:28:43 2022
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
X-Patchwork-Submitter: Daniel Golle <daniel@makrotopia.org>
X-Patchwork-Id: 12979251
X-Patchwork-Delegate: kvalo@adurom.com
Return-Path: <linux-wireless-owner@kernel.org>
Date: Sat, 17 Sep 2022 21:28:43 +0100
From: Daniel Golle <daniel@makrotopia.org>
To: linux-wireless@vger.kernel.org, Stanislaw Gruszka <stf_xl@wp.pl>,
Helmut Schaa <helmut.schaa@googlemail.com>
Cc: Kalle Valo <kvalo@kernel.org>,
"David S. Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@google.com>,
Jakub Kicinski <kuba@kernel.org>,
Paolo Abeni <pabeni@redhat.com>,
Johannes Berg <johannes.berg@intel.com>
Subject: [PATCH v3 10/16] rt2x00: add TX LOFT calibration for MT7620
Message-ID:
<d9295a9138a1f552b648aacb84e1419d38f5c896.1663445157.git.daniel@makrotopia.org>
References: <cover.1663445157.git.daniel@makrotopia.org>
MIME-Version: 1.0
Content-Disposition: inline
In-Reply-To: <cover.1663445157.git.daniel@makrotopia.org>
Precedence: bulk
List-ID: <linux-wireless.vger.kernel.org>
X-Mailing-List: linux-wireless@vger.kernel.org
From: Tomislav Požega <pozega.tomislav@gmail.com>
Add TX LOFT calibration from mtk driver.
Signed-off-by: Tomislav Požega <pozega.tomislav@gmail.com>
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
---
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
v2: use helper functions, make tables static const, remove useless
debug prints
v3: don't export function not used anywhere else
Reported-by: kernel test robot <lkp@intel.com>
.../net/wireless/ralink/rt2x00/rt2800lib.c | 902 ++++++++++++++++++
.../net/wireless/ralink/rt2x00/rt2800lib.h | 10 +
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
2 files changed, 912 insertions(+)
--- a/drivers/net/wireless/ralink/rt2x00/rt2800lib.c
+++ b/drivers/net/wireless/ralink/rt2x00/rt2800lib.c
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
@@ -9041,6 +9041,907 @@ restore_value:
rt2800_register_write(rt2x00dev, MAC_SYS_CTRL, savemacsysctrl);
}
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+static void rt2800_rf_configstore(struct rt2x00_dev *rt2x00dev,
+ struct rf_reg_pair rf_reg_record[][13], u8 chain)
+{
+ u8 rfvalue = 0;
+
+ if (chain == CHAIN_0) {
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 0, 1);
+ rf_reg_record[CHAIN_0][0].bank = 0;
+ rf_reg_record[CHAIN_0][0].reg = 1;
+ rf_reg_record[CHAIN_0][0].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 0, 2);
+ rf_reg_record[CHAIN_0][1].bank = 0;
+ rf_reg_record[CHAIN_0][1].reg = 2;
+ rf_reg_record[CHAIN_0][1].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 0, 35);
+ rf_reg_record[CHAIN_0][2].bank = 0;
+ rf_reg_record[CHAIN_0][2].reg = 35;
+ rf_reg_record[CHAIN_0][2].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 0, 42);
+ rf_reg_record[CHAIN_0][3].bank = 0;
+ rf_reg_record[CHAIN_0][3].reg = 42;
+ rf_reg_record[CHAIN_0][3].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 4, 0);
+ rf_reg_record[CHAIN_0][4].bank = 4;
+ rf_reg_record[CHAIN_0][4].reg = 0;
+ rf_reg_record[CHAIN_0][4].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 4, 2);
+ rf_reg_record[CHAIN_0][5].bank = 4;
+ rf_reg_record[CHAIN_0][5].reg = 2;
+ rf_reg_record[CHAIN_0][5].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 4, 34);
+ rf_reg_record[CHAIN_0][6].bank = 4;
+ rf_reg_record[CHAIN_0][6].reg = 34;
+ rf_reg_record[CHAIN_0][6].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 5, 3);
+ rf_reg_record[CHAIN_0][7].bank = 5;
+ rf_reg_record[CHAIN_0][7].reg = 3;
+ rf_reg_record[CHAIN_0][7].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 5, 4);
+ rf_reg_record[CHAIN_0][8].bank = 5;
+ rf_reg_record[CHAIN_0][8].reg = 4;
+ rf_reg_record[CHAIN_0][8].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 5, 17);
+ rf_reg_record[CHAIN_0][9].bank = 5;
+ rf_reg_record[CHAIN_0][9].reg = 17;
+ rf_reg_record[CHAIN_0][9].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 5, 18);
+ rf_reg_record[CHAIN_0][10].bank = 5;
+ rf_reg_record[CHAIN_0][10].reg = 18;
+ rf_reg_record[CHAIN_0][10].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 5, 19);
+ rf_reg_record[CHAIN_0][11].bank = 5;
+ rf_reg_record[CHAIN_0][11].reg = 19;
+ rf_reg_record[CHAIN_0][11].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 5, 20);
+ rf_reg_record[CHAIN_0][12].bank = 5;
+ rf_reg_record[CHAIN_0][12].reg = 20;
+ rf_reg_record[CHAIN_0][12].value = rfvalue;
+ } else if (chain == CHAIN_1) {
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 0, 1);
+ rf_reg_record[CHAIN_1][0].bank = 0;
+ rf_reg_record[CHAIN_1][0].reg = 1;
+ rf_reg_record[CHAIN_1][0].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 0, 2);
+ rf_reg_record[CHAIN_1][1].bank = 0;
+ rf_reg_record[CHAIN_1][1].reg = 2;
+ rf_reg_record[CHAIN_1][1].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 0, 35);
+ rf_reg_record[CHAIN_1][2].bank = 0;
+ rf_reg_record[CHAIN_1][2].reg = 35;
+ rf_reg_record[CHAIN_1][2].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 0, 42);
+ rf_reg_record[CHAIN_1][3].bank = 0;
+ rf_reg_record[CHAIN_1][3].reg = 42;
+ rf_reg_record[CHAIN_1][3].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 6, 0);
+ rf_reg_record[CHAIN_1][4].bank = 6;
+ rf_reg_record[CHAIN_1][4].reg = 0;
+ rf_reg_record[CHAIN_1][4].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 6, 2);
+ rf_reg_record[CHAIN_1][5].bank = 6;
+ rf_reg_record[CHAIN_1][5].reg = 2;
+ rf_reg_record[CHAIN_1][5].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 6, 34);
+ rf_reg_record[CHAIN_1][6].bank = 6;
+ rf_reg_record[CHAIN_1][6].reg = 34;
+ rf_reg_record[CHAIN_1][6].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 7, 3);
+ rf_reg_record[CHAIN_1][7].bank = 7;
+ rf_reg_record[CHAIN_1][7].reg = 3;
+ rf_reg_record[CHAIN_1][7].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 7, 4);
+ rf_reg_record[CHAIN_1][8].bank = 7;
+ rf_reg_record[CHAIN_1][8].reg = 4;
+ rf_reg_record[CHAIN_1][8].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 7, 17);
+ rf_reg_record[CHAIN_1][9].bank = 7;
+ rf_reg_record[CHAIN_1][9].reg = 17;
+ rf_reg_record[CHAIN_1][9].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 7, 18);
+ rf_reg_record[CHAIN_1][10].bank = 7;
+ rf_reg_record[CHAIN_1][10].reg = 18;
+ rf_reg_record[CHAIN_1][10].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 7, 19);
+ rf_reg_record[CHAIN_1][11].bank = 7;
+ rf_reg_record[CHAIN_1][11].reg = 19;
+ rf_reg_record[CHAIN_1][11].value = rfvalue;
+ rfvalue = rt2800_rfcsr_read_bank(rt2x00dev, 7, 20);
+ rf_reg_record[CHAIN_1][12].bank = 7;
+ rf_reg_record[CHAIN_1][12].reg = 20;
+ rf_reg_record[CHAIN_1][12].value = rfvalue;
+ } else {
+ rt2x00_warn(rt2x00dev, "Unknown chain = %u\n", chain);
+ }
+}
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+static void rt2800_rf_configrecover(struct rt2x00_dev *rt2x00dev,
+ struct rf_reg_pair rf_record[][13])
+{
+ u8 chain_index = 0, record_index = 0;
+ u8 bank = 0, rf_register = 0, value = 0;
+
+ for (chain_index = 0; chain_index < 2; chain_index++) {
+ for (record_index = 0; record_index < 13; record_index++) {
+ bank = rf_record[chain_index][record_index].bank;
+ rf_register = rf_record[chain_index][record_index].reg;
+ value = rf_record[chain_index][record_index].value;
+ rt2800_rfcsr_write_bank(rt2x00dev, bank, rf_register, value);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ rt2x00_dbg(rt2x00dev, "bank: %d, rf_register: %d, value: %x\n",
+ bank, rf_register, value);
+ }
+ }
+}
+
+static void rt2800_setbbptonegenerator(struct rt2x00_dev *rt2x00dev)
+{
+ rt2800_bbp_write(rt2x00dev, 158, 0xAA);
+ rt2800_bbp_write(rt2x00dev, 159, 0x00);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0xAB);
+ rt2800_bbp_write(rt2x00dev, 159, 0x0A);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0xAC);
+ rt2800_bbp_write(rt2x00dev, 159, 0x3F);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0xAD);
+ rt2800_bbp_write(rt2x00dev, 159, 0x3F);
+
+ rt2800_bbp_write(rt2x00dev, 244, 0x40);
+}
+
+static u32 rt2800_do_fft_accumulation(struct rt2x00_dev *rt2x00dev, u8 tidx, u8 read_neg)
+{
+ u32 macvalue = 0;
+ int fftout_i = 0, fftout_q = 0;
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ u32 ptmp = 0, pint = 0;
+ u8 bbp = 0;
+ u8 tidxi;
+
+ rt2800_bbp_write(rt2x00dev, 158, 0x00);
+ rt2800_bbp_write(rt2x00dev, 159, 0x9b);
+
+ bbp = 0x9b;
+
+ while (bbp == 0x9b) {
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ usleep_range(10, 50);
+ bbp = rt2800_bbp_read(rt2x00dev, 159);
+ bbp = bbp & 0xff;
+ }
+
+ rt2800_bbp_write(rt2x00dev, 158, 0xba);
+ rt2800_bbp_write(rt2x00dev, 159, tidx);
+ rt2800_bbp_write(rt2x00dev, 159, tidx);
+ rt2800_bbp_write(rt2x00dev, 159, tidx);
+
+ macvalue = rt2800_register_read(rt2x00dev, 0x057C);
+
+ fftout_i = (macvalue >> 16);
+ fftout_i = (fftout_i & 0x8000) ? (fftout_i - 0x10000) : fftout_i;
+ fftout_q = (macvalue & 0xffff);
+ fftout_q = (fftout_q & 0x8000) ? (fftout_q - 0x10000) : fftout_q;
+ ptmp = (fftout_i * fftout_i);
+ ptmp = ptmp + (fftout_q * fftout_q);
+ pint = ptmp;
+ rt2x00_dbg(rt2x00dev, "I = %d, Q = %d, power = %x\n", fftout_i, fftout_q, pint);
+ if (read_neg) {
+ pint = pint >> 1;
+ tidxi = 0x40 - tidx;
+ tidxi = tidxi & 0x3f;
+
+ rt2800_bbp_write(rt2x00dev, 158, 0xba);
+ rt2800_bbp_write(rt2x00dev, 159, tidxi);
+ rt2800_bbp_write(rt2x00dev, 159, tidxi);
+ rt2800_bbp_write(rt2x00dev, 159, tidxi);
+
+ macvalue = rt2800_register_read(rt2x00dev, 0x057C);
+
+ fftout_i = (macvalue >> 16);
+ fftout_i = (fftout_i & 0x8000) ? (fftout_i - 0x10000) : fftout_i;
+ fftout_q = (macvalue & 0xffff);
+ fftout_q = (fftout_q & 0x8000) ? (fftout_q - 0x10000) : fftout_q;
+ ptmp = (fftout_i * fftout_i);
+ ptmp = ptmp + (fftout_q * fftout_q);
+ ptmp = ptmp >> 1;
+ pint = pint + ptmp;
+ }
+
+ return pint;
+}
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+static u32 rt2800_read_fft_accumulation(struct rt2x00_dev *rt2x00dev, u8 tidx)
+{
+ u32 macvalue = 0;
+ int fftout_i = 0, fftout_q = 0;
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ u32 ptmp = 0, pint = 0;
+
+ rt2800_bbp_write(rt2x00dev, 158, 0xBA);
+ rt2800_bbp_write(rt2x00dev, 159, tidx);
+ rt2800_bbp_write(rt2x00dev, 159, tidx);
+ rt2800_bbp_write(rt2x00dev, 159, tidx);
+
+ macvalue = rt2800_register_read(rt2x00dev, 0x057C);
+
+ fftout_i = (macvalue >> 16);
+ fftout_i = (fftout_i & 0x8000) ? (fftout_i - 0x10000) : fftout_i;
+ fftout_q = (macvalue & 0xffff);
+ fftout_q = (fftout_q & 0x8000) ? (fftout_q - 0x10000) : fftout_q;
+ ptmp = (fftout_i * fftout_i);
+ ptmp = ptmp + (fftout_q * fftout_q);
+ pint = ptmp;
+
+ return pint;
+}
+
+static void rt2800_write_dc(struct rt2x00_dev *rt2x00dev, u8 ch_idx, u8 alc, u8 iorq, u8 dc)
+{
+ u8 bbp = 0;
+
+ rt2800_bbp_write(rt2x00dev, 158, 0xb0);
+ bbp = alc | 0x80;
+ rt2800_bbp_write(rt2x00dev, 159, bbp);
+
+ if (ch_idx == 0)
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ bbp = (iorq == 0) ? 0xb1 : 0xb2;
+ else
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ bbp = (iorq == 0) ? 0xb8 : 0xb9;
+
+ rt2800_bbp_write(rt2x00dev, 158, bbp);
+ bbp = dc;
+ rt2800_bbp_write(rt2x00dev, 159, bbp);
+}
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+static void rt2800_loft_search(struct rt2x00_dev *rt2x00dev, u8 ch_idx,
+ u8 alc_idx, u8 dc_result[][RF_ALC_NUM][2])
+{
+ u32 p0 = 0, p1 = 0, pf = 0;
+ char idx0 = 0, idx1 = 0;
+ u8 idxf[] = {0x00, 0x00};
+ u8 ibit = 0x20;
+ u8 iorq;
+ char bidx;
+
+ rt2800_bbp_write(rt2x00dev, 158, 0xb0);
+ rt2800_bbp_write(rt2x00dev, 159, 0x80);
+
+ for (bidx = 5; bidx >= 0; bidx--) {
+ for (iorq = 0; iorq <= 1; iorq++) {
+ if (idxf[iorq] == 0x20) {
+ idx0 = 0x20;
+ p0 = pf;
+ } else {
+ idx0 = idxf[iorq] - ibit;
+ idx0 = idx0 & 0x3F;
+ rt2800_write_dc(rt2x00dev, ch_idx, 0, iorq, idx0);
+ p0 = rt2800_do_fft_accumulation(rt2x00dev, 0x0A, 0);
+ }
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ idx1 = idxf[iorq] + (bidx == 5 ? 0 : ibit);
+ idx1 = idx1 & 0x3F;
+ rt2800_write_dc(rt2x00dev, ch_idx, 0, iorq, idx1);
+ p1 = rt2800_do_fft_accumulation(rt2x00dev, 0x0A, 0);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ rt2x00_dbg(rt2x00dev, "alc=%u, IorQ=%u, idx_final=%2x\n",
+ alc_idx, iorq, idxf[iorq]);
+ rt2x00_dbg(rt2x00dev, "p0=%x, p1=%x, pf=%x, idx_0=%x, idx_1=%x, ibit=%x\n",
+ p0, p1, pf, idx0, idx1, ibit);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if (bidx != 5 && pf <= p0 && pf < p1) {
+ idxf[iorq] = idxf[iorq];
+ } else if (p0 < p1) {
+ pf = p0;
+ idxf[iorq] = idx0 & 0x3F;
+ } else {
+ pf = p1;
+ idxf[iorq] = idx1 & 0x3F;
+ }
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ rt2x00_dbg(rt2x00dev, "IorQ=%u, idx_final[%u]:%x, pf:%8x\n",
+ iorq, iorq, idxf[iorq], pf);
+
+ rt2800_write_dc(rt2x00dev, ch_idx, 0, iorq, idxf[iorq]);
+ }
+ ibit = ibit >> 1;
+ }
+ dc_result[ch_idx][alc_idx][0] = idxf[0];
+ dc_result[ch_idx][alc_idx][1] = idxf[1];
+}
+
+static void rt2800_iq_search(struct rt2x00_dev *rt2x00dev, u8 ch_idx, u8 *ges, u8 *pes)
+{
+ u32 p0 = 0, p1 = 0, pf = 0;
+ char perr = 0, gerr = 0, iq_err = 0;
+ char pef = 0, gef = 0;
+ char psta, pend;
+ char gsta, gend;
+
+ u8 ibit = 0x20;
+ u8 first_search = 0x00, touch_neg_max = 0x00;
+ char idx0 = 0, idx1 = 0;
+ u8 gop;
+ u8 bbp = 0;
+ char bidx;
+
+ for (bidx = 5; bidx >= 1; bidx--) {
+ for (gop = 0; gop < 2; gop++) {
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if (gop == 1 || bidx < 4) {
+ if (gop == 0)
+ iq_err = gerr;
+ else
+ iq_err = perr;
+
+ first_search = (gop == 0) ? (bidx == 3) : (bidx == 5);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ touch_neg_max = (gop) ? ((iq_err & 0x0F) == 0x08) :
+ ((iq_err & 0x3F) == 0x20);
+
+ if (touch_neg_max) {
+ p0 = pf;
+ idx0 = iq_err;
+ } else {
+ idx0 = iq_err - ibit;
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ bbp = (ch_idx == 0) ? ((gop == 0) ? 0x28 : 0x29) :
+ ((gop == 0) ? 0x46 : 0x47);
+
+ rt2800_bbp_write(rt2x00dev, 158, bbp);
+ rt2800_bbp_write(rt2x00dev, 159, idx0);
+
+ p0 = rt2800_do_fft_accumulation(rt2x00dev, 0x14, 1);
+ }
+
+ idx1 = iq_err + (first_search ? 0 : ibit);
+ idx1 = (gop == 0) ? (idx1 & 0x0F) : (idx1 & 0x3F);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ bbp = (ch_idx == 0) ? (gop == 0) ? 0x28 : 0x29 :
+ (gop == 0) ? 0x46 : 0x47;
+
+ rt2800_bbp_write(rt2x00dev, 158, bbp);
+ rt2800_bbp_write(rt2x00dev, 159, idx1);
+
+ p1 = rt2800_do_fft_accumulation(rt2x00dev, 0x14, 1);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ rt2x00_dbg(rt2x00dev,
+ "p0=%x, p1=%x, pwer_final=%x, idx0=%x, idx1=%x, iq_err=%x, gop=%d, ibit=%x\n",
+ p0, p1, pf, idx0, idx1, iq_err, gop, ibit);
+
+ if (!(!first_search && pf <= p0 && pf < p1)) {
+ if (p0 < p1) {
+ pf = p0;
+ iq_err = idx0;
+ } else {
+ pf = p1;
+ iq_err = idx1;
+ }
+ }
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ bbp = (ch_idx == 0) ? (gop == 0) ? 0x28 : 0x29 :
+ (gop == 0) ? 0x46 : 0x47;
+
+ rt2800_bbp_write(rt2x00dev, 158, bbp);
+ rt2800_bbp_write(rt2x00dev, 159, iq_err);
+
+ if (gop == 0)
+ gerr = iq_err;
+ else
+ perr = iq_err;
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ rt2x00_dbg(rt2x00dev, "IQCalibration pf=%8x (%2x, %2x) !\n",
+ pf, gerr & 0x0F, perr & 0x3F);
+ }
+ }
+
+ if (bidx > 0)
+ ibit = (ibit >> 1);
+ }
+ gerr = (gerr & 0x08) ? (gerr & 0x0F) - 0x10 : (gerr & 0x0F);
+ perr = (perr & 0x20) ? (perr & 0x3F) - 0x40 : (perr & 0x3F);
+
+ gerr = (gerr < -0x07) ? -0x07 : (gerr > 0x05) ? 0x05 : gerr;
+ gsta = gerr - 1;
+ gend = gerr + 2;
+
+ perr = (perr < -0x1f) ? -0x1f : (perr > 0x1d) ? 0x1d : perr;
+ psta = perr - 1;
+ pend = perr + 2;
+
+ for (gef = gsta; gef <= gend; gef = gef + 1)
+ for (pef = psta; pef <= pend; pef = pef + 1) {
+ bbp = (ch_idx == 0) ? 0x28 : 0x46;
+ rt2800_bbp_write(rt2x00dev, 158, bbp);
+ rt2800_bbp_write(rt2x00dev, 159, gef & 0x0F);
+
+ bbp = (ch_idx == 0) ? 0x29 : 0x47;
+ rt2800_bbp_write(rt2x00dev, 158, bbp);
+ rt2800_bbp_write(rt2x00dev, 159, pef & 0x3F);
+
+ p1 = rt2800_do_fft_accumulation(rt2x00dev, 0x14, 1);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if (gef == gsta && pef == psta) {
+ pf = p1;
+ gerr = gef;
+ perr = pef;
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ } else if (pf > p1) {
+ pf = p1;
+ gerr = gef;
+ perr = pef;
+ }
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ rt2x00_dbg(rt2x00dev, "Fine IQCalibration p1=%8x pf=%8x (%2x, %2x) !\n",
+ p1, pf, gef & 0x0F, pef & 0x3F);
+ }
+
+ ges[ch_idx] = gerr & 0x0F;
+ pes[ch_idx] = perr & 0x3F;
+}
+
+static void rt2800_rf_aux_tx0_loopback(struct rt2x00_dev *rt2x00dev)
+{
+ rt2800_rfcsr_write_bank(rt2x00dev, 0, 1, 0x21);
+ rt2800_rfcsr_write_bank(rt2x00dev, 0, 2, 0x10);
+ rt2800_rfcsr_write_bank(rt2x00dev, 0, 35, 0x00);
+ rt2800_rfcsr_write_bank(rt2x00dev, 0, 42, 0x1b);
+ rt2800_rfcsr_write_bank(rt2x00dev, 4, 0, 0x81);
+ rt2800_rfcsr_write_bank(rt2x00dev, 4, 2, 0x81);
+ rt2800_rfcsr_write_bank(rt2x00dev, 4, 34, 0xee);
+ rt2800_rfcsr_write_bank(rt2x00dev, 5, 3, 0x2d);
+ rt2800_rfcsr_write_bank(rt2x00dev, 5, 4, 0x2d);
+ rt2800_rfcsr_write_bank(rt2x00dev, 5, 17, 0x80);
+ rt2800_rfcsr_write_bank(rt2x00dev, 5, 18, 0xd7);
+ rt2800_rfcsr_write_bank(rt2x00dev, 5, 19, 0xa2);
+ rt2800_rfcsr_write_bank(rt2x00dev, 5, 20, 0x20);
+}
+
+static void rt2800_rf_aux_tx1_loopback(struct rt2x00_dev *rt2x00dev)
+{
+ rt2800_rfcsr_write_bank(rt2x00dev, 0, 1, 0x22);
+ rt2800_rfcsr_write_bank(rt2x00dev, 0, 2, 0x20);
+ rt2800_rfcsr_write_bank(rt2x00dev, 0, 35, 0x00);
+ rt2800_rfcsr_write_bank(rt2x00dev, 0, 42, 0x4b);
+ rt2800_rfcsr_write_bank(rt2x00dev, 6, 0, 0x81);
+ rt2800_rfcsr_write_bank(rt2x00dev, 6, 2, 0x81);
+ rt2800_rfcsr_write_bank(rt2x00dev, 6, 34, 0xee);
+ rt2800_rfcsr_write_bank(rt2x00dev, 7, 3, 0x2d);
+ rt2800_rfcsr_write_bank(rt2x00dev, 7, 4, 0x2d);
+ rt2800_rfcsr_write_bank(rt2x00dev, 7, 17, 0x80);
+ rt2800_rfcsr_write_bank(rt2x00dev, 7, 18, 0xd7);
+ rt2800_rfcsr_write_bank(rt2x00dev, 7, 19, 0xa2);
+ rt2800_rfcsr_write_bank(rt2x00dev, 7, 20, 0x20);
+}
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+static void rt2800_loft_iq_calibration(struct rt2x00_dev *rt2x00dev)
+{
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ struct rf_reg_pair rf_store[CHAIN_NUM][13];
+ u32 macorg1 = 0;
+ u32 macorg2 = 0;
+ u32 macorg3 = 0;
+ u32 macorg4 = 0;
+ u32 macorg5 = 0;
+ u32 orig528 = 0;
+ u32 orig52c = 0;
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ u32 savemacsysctrl = 0;
+ u32 macvalue = 0;
+ u32 mac13b8 = 0;
+ u32 p0 = 0, p1 = 0;
+ u32 p0_idx10 = 0, p1_idx10 = 0;
+
+ u8 rfvalue;
+ u8 loft_dc_search_result[CHAIN_NUM][RF_ALC_NUM][2];
+ u8 ger[CHAIN_NUM], per[CHAIN_NUM];
+
+ u8 vga_gain[] = {14, 14};
+ u8 bbp = 0, ch_idx = 0, rf_alc_idx = 0, idx = 0;
+ u8 bbpr30, rfb0r39, rfb0r42;
+ u8 bbpr1;
+ u8 bbpr4;
+ u8 bbpr241, bbpr242;
+ u8 count_step;
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ static const u8 rf_gain[] = {0x00, 0x01, 0x02, 0x04, 0x08, 0x0c};
+ static const u8 rfvga_gain_table[] = {0x24, 0x25, 0x26, 0x27, 0x28, 0x2c, 0x2d, 0x2e, 0x2f, 0x30,
+ 0x31, 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3F};
+ static const u8 bbp_2324gain[] = {0x16, 0x14, 0x12, 0x10, 0x0c, 0x08};
+
+ savemacsysctrl = rt2800_register_read(rt2x00dev, MAC_SYS_CTRL);
+ macorg1 = rt2800_register_read(rt2x00dev, TX_PIN_CFG);
+ macorg2 = rt2800_register_read(rt2x00dev, RF_CONTROL0);
+ macorg3 = rt2800_register_read(rt2x00dev, RF_BYPASS0);
+ macorg4 = rt2800_register_read(rt2x00dev, RF_CONTROL3);
+ macorg5 = rt2800_register_read(rt2x00dev, RF_BYPASS3);
+ mac13b8 = rt2800_register_read(rt2x00dev, 0x13b8);
+ orig528 = rt2800_register_read(rt2x00dev, RF_CONTROL2);
+ orig52c = rt2800_register_read(rt2x00dev, RF_BYPASS2);
+
+ macvalue = rt2800_register_read(rt2x00dev, MAC_SYS_CTRL);
+ macvalue &= (~0x04);
+ rt2800_register_write(rt2x00dev, MAC_SYS_CTRL, macvalue);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if (unlikely(rt2800_wait_bbp_rf_ready(rt2x00dev, MAC_STATUS_CFG_BBP_RF_BUSY_TX)))
+ rt2x00_warn(rt2x00dev, "RF TX busy in LOFT IQ calibration\n");
+
+ macvalue = rt2800_register_read(rt2x00dev, MAC_SYS_CTRL);
+ macvalue &= (~0x08);
+ rt2800_register_write(rt2x00dev, MAC_SYS_CTRL, macvalue);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if (unlikely(rt2800_wait_bbp_rf_ready(rt2x00dev, MAC_STATUS_CFG_BBP_RF_BUSY_RX)))
+ rt2x00_warn(rt2x00dev, "RF RX busy in LOFT IQ calibration\n");
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ for (ch_idx = 0; ch_idx < 2; ch_idx++)
+ rt2800_rf_configstore(rt2x00dev, rf_store, ch_idx);
+
+ bbpr30 = rt2800_bbp_read(rt2x00dev, 30);
+ rfb0r39 = rt2800_rfcsr_read_bank(rt2x00dev, 0, 39);
+ rfb0r42 = rt2800_rfcsr_read_bank(rt2x00dev, 0, 42);
+
+ rt2800_bbp_write(rt2x00dev, 30, 0x1F);
+ rt2800_rfcsr_write_bank(rt2x00dev, 0, 39, 0x80);
+ rt2800_rfcsr_write_bank(rt2x00dev, 0, 42, 0x5B);
+
+ rt2800_bbp_write(rt2x00dev, 23, 0x00);
+ rt2800_bbp_write(rt2x00dev, 24, 0x00);
+
+ rt2800_setbbptonegenerator(rt2x00dev);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ for (ch_idx = 0; ch_idx < 2; ch_idx++) {
+ rt2800_bbp_write(rt2x00dev, 23, 0x00);
+ rt2800_bbp_write(rt2x00dev, 24, 0x00);
+ rt2800_register_write(rt2x00dev, MAC_SYS_CTRL, 0x00);
+ rt2800_register_write(rt2x00dev, TX_PIN_CFG, 0x0000000F);
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, 0x00000004);
+ rt2800_register_write(rt2x00dev, RF_BYPASS0, 0x00003306);
+ rt2800_register_write(rt2x00dev, 0x13b8, 0x10);
+ udelay(1);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if (ch_idx == 0)
+ rt2800_rf_aux_tx0_loopback(rt2x00dev);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ else
+ rt2800_rf_aux_tx1_loopback(rt2x00dev);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+
+ udelay(1);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if (ch_idx == 0)
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, 0x00001004);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ else
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, 0x00002004);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0x05);
+ rt2800_bbp_write(rt2x00dev, 159, 0x00);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0x01);
+ if (ch_idx == 0)
+ rt2800_bbp_write(rt2x00dev, 159, 0x00);
+ else
+ rt2800_bbp_write(rt2x00dev, 159, 0x01);
+
+ vga_gain[ch_idx] = 18;
+ for (rf_alc_idx = 0; rf_alc_idx < 3; rf_alc_idx++) {
+ rt2800_bbp_write(rt2x00dev, 23, bbp_2324gain[rf_alc_idx]);
+ rt2800_bbp_write(rt2x00dev, 24, bbp_2324gain[rf_alc_idx]);
+
+ macvalue = rt2800_register_read(rt2x00dev, RF_CONTROL3);
+ macvalue &= (~0x0000F1F1);
+ macvalue |= (rf_gain[rf_alc_idx] << 4);
+ macvalue |= (rf_gain[rf_alc_idx] << 12);
+ rt2800_register_write(rt2x00dev, RF_CONTROL3, macvalue);
+ macvalue = (0x0000F1F1);
+ rt2800_register_write(rt2x00dev, RF_BYPASS3, macvalue);
+
+ if (rf_alc_idx == 0) {
+ rt2800_write_dc(rt2x00dev, ch_idx, 0, 1, 0x21);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ for (; vga_gain[ch_idx] > 0;
+ vga_gain[ch_idx] = vga_gain[ch_idx] - 2) {
+ rfvalue = rfvga_gain_table[vga_gain[ch_idx]];
+ rt2800_rfcsr_write_dccal(rt2x00dev, 3, rfvalue);
+ rt2800_rfcsr_write_dccal(rt2x00dev, 4, rfvalue);
+ rt2800_write_dc(rt2x00dev, ch_idx, 0, 1, 0x00);
+ rt2800_write_dc(rt2x00dev, ch_idx, 0, 0, 0x00);
+ p0 = rt2800_do_fft_accumulation(rt2x00dev, 0x0A, 0);
+ rt2800_write_dc(rt2x00dev, ch_idx, 0, 0, 0x21);
+ p1 = rt2800_do_fft_accumulation(rt2x00dev, 0x0A, 0);
+ rt2x00_dbg(rt2x00dev, "LOFT AGC %d %d\n", p0, p1);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if ((p0 < 7000 * 7000) && (p1 < (7000 * 7000)))
+ break;
+ }
+
+ rt2800_write_dc(rt2x00dev, ch_idx, 0, 0, 0x00);
+ rt2800_write_dc(rt2x00dev, ch_idx, 0, 1, 0x00);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ rt2x00_dbg(rt2x00dev, "Used VGA %d %x\n", vga_gain[ch_idx],
+ rfvga_gain_table[vga_gain[ch_idx]]);
+
+ if (vga_gain[ch_idx] < 0)
+ vga_gain[ch_idx] = 0;
+ }
+
+ rfvalue = rfvga_gain_table[vga_gain[ch_idx]];
+
+ rt2800_rfcsr_write_dccal(rt2x00dev, 3, rfvalue);
+ rt2800_rfcsr_write_dccal(rt2x00dev, 4, rfvalue);
+
+ rt2800_loft_search(rt2x00dev, ch_idx, rf_alc_idx, loft_dc_search_result);
+ }
+ }
+
+ for (rf_alc_idx = 0; rf_alc_idx < 3; rf_alc_idx++) {
+ for (idx = 0; idx < 4; idx++) {
+ rt2800_bbp_write(rt2x00dev, 158, 0xB0);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ bbp = (idx << 2) + rf_alc_idx;
+ rt2800_bbp_write(rt2x00dev, 159, bbp);
+ rt2x00_dbg(rt2x00dev, " ALC %2x,", bbp);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0xb1);
+ bbp = loft_dc_search_result[CHAIN_0][rf_alc_idx][0x00];
+ bbp = bbp & 0x3F;
+ rt2800_bbp_write(rt2x00dev, 159, bbp);
+ rt2x00_dbg(rt2x00dev, " I0 %2x,", bbp);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0xb2);
+ bbp = loft_dc_search_result[CHAIN_0][rf_alc_idx][0x01];
+ bbp = bbp & 0x3F;
+ rt2800_bbp_write(rt2x00dev, 159, bbp);
+ rt2x00_dbg(rt2x00dev, " Q0 %2x,", bbp);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0xb8);
+ bbp = loft_dc_search_result[CHAIN_1][rf_alc_idx][0x00];
+ bbp = bbp & 0x3F;
+ rt2800_bbp_write(rt2x00dev, 159, bbp);
+ rt2x00_dbg(rt2x00dev, " I1 %2x,", bbp);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0xb9);
+ bbp = loft_dc_search_result[CHAIN_1][rf_alc_idx][0x01];
+ bbp = bbp & 0x3F;
+ rt2800_bbp_write(rt2x00dev, 159, bbp);
+ rt2x00_dbg(rt2x00dev, " Q1 %2x\n", bbp);
+ }
+ }
+
+ rt2800_bbp_write(rt2x00dev, 23, 0x00);
+ rt2800_bbp_write(rt2x00dev, 24, 0x00);
+
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, 0x04);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0x00);
+ rt2800_bbp_write(rt2x00dev, 159, 0x00);
+
+ bbp = 0x00;
+ rt2800_bbp_write(rt2x00dev, 244, 0x00);
+
+ rt2800_bbp_write(rt2x00dev, 21, 0x01);
+ udelay(1);
+ rt2800_bbp_write(rt2x00dev, 21, 0x00);
+
+ rt2800_rf_configrecover(rt2x00dev, rf_store);
+
+ rt2800_register_write(rt2x00dev, TX_PIN_CFG, macorg1);
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, 0x04);
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, 0x00);
+ rt2800_register_write(rt2x00dev, RF_BYPASS0, 0x00);
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, macorg2);
+ udelay(1);
+ rt2800_register_write(rt2x00dev, RF_BYPASS0, macorg3);
+ rt2800_register_write(rt2x00dev, RF_CONTROL3, macorg4);
+ rt2800_register_write(rt2x00dev, RF_BYPASS3, macorg5);
+ rt2800_register_write(rt2x00dev, MAC_SYS_CTRL, savemacsysctrl);
+ rt2800_register_write(rt2x00dev, RF_CONTROL2, orig528);
+ rt2800_register_write(rt2x00dev, RF_BYPASS2, orig52c);
+ rt2800_register_write(rt2x00dev, 0x13b8, mac13b8);
+
+ savemacsysctrl = rt2800_register_read(rt2x00dev, MAC_SYS_CTRL);
+ macorg1 = rt2800_register_read(rt2x00dev, TX_PIN_CFG);
+ macorg2 = rt2800_register_read(rt2x00dev, RF_CONTROL0);
+ macorg3 = rt2800_register_read(rt2x00dev, RF_BYPASS0);
+ macorg4 = rt2800_register_read(rt2x00dev, RF_CONTROL3);
+ macorg5 = rt2800_register_read(rt2x00dev, RF_BYPASS3);
+
+ bbpr1 = rt2800_bbp_read(rt2x00dev, 1);
+ bbpr4 = rt2800_bbp_read(rt2x00dev, 4);
+ bbpr241 = rt2800_bbp_read(rt2x00dev, 241);
+ bbpr242 = rt2800_bbp_read(rt2x00dev, 242);
+ mac13b8 = rt2800_register_read(rt2x00dev, 0x13b8);
+
+ macvalue = rt2800_register_read(rt2x00dev, MAC_SYS_CTRL);
+ macvalue &= (~0x04);
+ rt2800_register_write(rt2x00dev, MAC_SYS_CTRL, macvalue);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+
+ if (unlikely(rt2800_wait_bbp_rf_ready(rt2x00dev, MAC_STATUS_CFG_BBP_RF_BUSY_TX)))
+ rt2x00_warn(rt2x00dev, "RF TX busy in LOFT IQ calibration\n");
+
+ macvalue = rt2800_register_read(rt2x00dev, MAC_SYS_CTRL);
+ macvalue &= (~0x08);
+ rt2800_register_write(rt2x00dev, MAC_SYS_CTRL, macvalue);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+
+ if (unlikely(rt2800_wait_bbp_rf_ready(rt2x00dev, MAC_STATUS_CFG_BBP_RF_BUSY_RX)))
+ rt2x00_warn(rt2x00dev, "RF RX busy in LOFT IQ calibration\n");
+
+ if (test_bit(CAPABILITY_EXTERNAL_PA_TX0, &rt2x00dev->cap_flags)) {
+ rt2800_register_write(rt2x00dev, RF_CONTROL3, 0x00000101);
+ rt2800_register_write(rt2x00dev, RF_BYPASS3, 0x0000F1F1);
+ }
+
+ rt2800_bbp_write(rt2x00dev, 23, 0x00);
+ rt2800_bbp_write(rt2x00dev, 24, 0x00);
+
+ if (test_bit(CAPABILITY_EXTERNAL_PA_TX0, &rt2x00dev->cap_flags)) {
+ rt2800_bbp_write(rt2x00dev, 4, bbpr4 & (~0x18));
+ rt2800_bbp_write(rt2x00dev, 21, 0x01);
+ udelay(1);
+ rt2800_bbp_write(rt2x00dev, 21, 0x00);
+
+ rt2800_bbp_write(rt2x00dev, 241, 0x14);
+ rt2800_bbp_write(rt2x00dev, 242, 0x80);
+ rt2800_bbp_write(rt2x00dev, 244, 0x31);
+ } else {
+ rt2800_setbbptonegenerator(rt2x00dev);
+ }
+
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, 0x00000004);
+ rt2800_register_write(rt2x00dev, RF_BYPASS0, 0x00003306);
+ udelay(1);
+
+ rt2800_register_write(rt2x00dev, TX_PIN_CFG, 0x0000000F);
+
+ if (!test_bit(CAPABILITY_EXTERNAL_PA_TX0, &rt2x00dev->cap_flags)) {
+ rt2800_register_write(rt2x00dev, RF_CONTROL3, 0x00000000);
+ rt2800_register_write(rt2x00dev, RF_BYPASS3, 0x0000F1F1);
+ }
+
+ rt2800_register_write(rt2x00dev, 0x13b8, 0x00000010);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ for (ch_idx = 0; ch_idx < 2; ch_idx++)
+ rt2800_rf_configstore(rt2x00dev, rf_store, ch_idx);
+
+ rt2800_rfcsr_write_dccal(rt2x00dev, 3, 0x3B);
+ rt2800_rfcsr_write_dccal(rt2x00dev, 4, 0x3B);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0x03);
+ rt2800_bbp_write(rt2x00dev, 159, 0x60);
+ rt2800_bbp_write(rt2x00dev, 158, 0xB0);
+ rt2800_bbp_write(rt2x00dev, 159, 0x80);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ for (ch_idx = 0; ch_idx < 2; ch_idx++) {
+ rt2800_bbp_write(rt2x00dev, 23, 0x00);
+ rt2800_bbp_write(rt2x00dev, 24, 0x00);
+
+ if (ch_idx == 0) {
+ rt2800_bbp_write(rt2x00dev, 158, 0x01);
+ rt2800_bbp_write(rt2x00dev, 159, 0x00);
+ if (test_bit(CAPABILITY_EXTERNAL_PA_TX0, &rt2x00dev->cap_flags)) {
+ bbp = bbpr1 & (~0x18);
+ bbp = bbp | 0x00;
+ rt2800_bbp_write(rt2x00dev, 1, bbp);
+ }
+ rt2800_rf_aux_tx0_loopback(rt2x00dev);
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, 0x00001004);
+ } else {
+ rt2800_bbp_write(rt2x00dev, 158, 0x01);
+ rt2800_bbp_write(rt2x00dev, 159, 0x01);
+ if (test_bit(CAPABILITY_EXTERNAL_PA_TX1, &rt2x00dev->cap_flags)) {
+ bbp = bbpr1 & (~0x18);
+ bbp = bbp | 0x08;
+ rt2800_bbp_write(rt2x00dev, 1, bbp);
+ }
+ rt2800_rf_aux_tx1_loopback(rt2x00dev);
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, 0x00002004);
+ }
+
+ rt2800_bbp_write(rt2x00dev, 158, 0x05);
+ rt2800_bbp_write(rt2x00dev, 159, 0x04);
+
+ bbp = (ch_idx == 0) ? 0x28 : 0x46;
+ rt2800_bbp_write(rt2x00dev, 158, bbp);
+ rt2800_bbp_write(rt2x00dev, 159, 0x00);
+
+ if (test_bit(CAPABILITY_EXTERNAL_PA_TX0, &rt2x00dev->cap_flags)) {
+ rt2800_bbp_write(rt2x00dev, 23, 0x06);
+ rt2800_bbp_write(rt2x00dev, 24, 0x06);
+ count_step = 1;
+ } else {
+ rt2800_bbp_write(rt2x00dev, 23, 0x1F);
+ rt2800_bbp_write(rt2x00dev, 24, 0x1F);
+ count_step = 2;
+ }
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ for (; vga_gain[ch_idx] < 19; vga_gain[ch_idx] = (vga_gain[ch_idx] + count_step)) {
+ rfvalue = rfvga_gain_table[vga_gain[ch_idx]];
+ rt2800_rfcsr_write_dccal(rt2x00dev, 3, rfvalue);
+ rt2800_rfcsr_write_dccal(rt2x00dev, 4, rfvalue);
+
+ bbp = (ch_idx == 0) ? 0x29 : 0x47;
+ rt2800_bbp_write(rt2x00dev, 158, bbp);
+ rt2800_bbp_write(rt2x00dev, 159, 0x00);
+ p0 = rt2800_do_fft_accumulation(rt2x00dev, 0x14, 0);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if (test_bit(CAPABILITY_EXTERNAL_PA_TX0, &rt2x00dev->cap_flags))
+ p0_idx10 = rt2800_read_fft_accumulation(rt2x00dev, 0x0A);
+
+ bbp = (ch_idx == 0) ? 0x29 : 0x47;
+ rt2800_bbp_write(rt2x00dev, 158, bbp);
+ rt2800_bbp_write(rt2x00dev, 159, 0x21);
+ p1 = rt2800_do_fft_accumulation(rt2x00dev, 0x14, 0);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if (test_bit(CAPABILITY_EXTERNAL_PA_TX1, &rt2x00dev->cap_flags))
+ p1_idx10 = rt2800_read_fft_accumulation(rt2x00dev, 0x0A);
+
+ rt2x00_dbg(rt2x00dev, "IQ AGC %d %d\n", p0, p1);
+
+ if (test_bit(CAPABILITY_EXTERNAL_PA_TX0, &rt2x00dev->cap_flags)) {
+ rt2x00_dbg(rt2x00dev, "IQ AGC IDX 10 %d %d\n", p0_idx10, p1_idx10);
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if ((p0_idx10 > 7000 * 7000) || (p1_idx10 > 7000 * 7000)) {
+ if (vga_gain[ch_idx] != 0)
+ vga_gain[ch_idx] = vga_gain[ch_idx] - 1;
+ break;
+ }
+ }
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if ((p0 > 2500 * 2500) || (p1 > 2500 * 2500))
+ break;
+ }
+
+ if (vga_gain[ch_idx] > 18)
+ vga_gain[ch_idx] = 18;
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ rt2x00_dbg(rt2x00dev, "Used VGA %d %x\n", vga_gain[ch_idx],
+ rfvga_gain_table[vga_gain[ch_idx]]);
+
+ bbp = (ch_idx == 0) ? 0x29 : 0x47;
+ rt2800_bbp_write(rt2x00dev, 158, bbp);
+ rt2800_bbp_write(rt2x00dev, 159, 0x00);
+
+ rt2800_iq_search(rt2x00dev, ch_idx, ger, per);
+ }
+
+ rt2800_bbp_write(rt2x00dev, 23, 0x00);
+ rt2800_bbp_write(rt2x00dev, 24, 0x00);
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, 0x04);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0x28);
+ bbp = ger[CHAIN_0] & 0x0F;
+ rt2800_bbp_write(rt2x00dev, 159, bbp);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0x29);
+ bbp = per[CHAIN_0] & 0x3F;
+ rt2800_bbp_write(rt2x00dev, 159, bbp);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0x46);
+ bbp = ger[CHAIN_1] & 0x0F;
+ rt2800_bbp_write(rt2x00dev, 159, bbp);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0x47);
+ bbp = per[CHAIN_1] & 0x3F;
+ rt2800_bbp_write(rt2x00dev, 159, bbp);
+
+ if (test_bit(CAPABILITY_EXTERNAL_PA_TX0, &rt2x00dev->cap_flags)) {
+ rt2800_bbp_write(rt2x00dev, 1, bbpr1);
+ rt2800_bbp_write(rt2x00dev, 241, bbpr241);
+ rt2800_bbp_write(rt2x00dev, 242, bbpr242);
+ }
+ rt2800_bbp_write(rt2x00dev, 244, 0x00);
+
+ rt2800_bbp_write(rt2x00dev, 158, 0x00);
+ rt2800_bbp_write(rt2x00dev, 159, 0x00);
+ rt2800_bbp_write(rt2x00dev, 158, 0xB0);
+ rt2800_bbp_write(rt2x00dev, 159, 0x00);
+
+ rt2800_bbp_write(rt2x00dev, 30, bbpr30);
+ rt2800_rfcsr_write_bank(rt2x00dev, 0, 39, rfb0r39);
+ rt2800_rfcsr_write_bank(rt2x00dev, 0, 42, rfb0r42);
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+ if (test_bit(CAPABILITY_EXTERNAL_PA_TX0, &rt2x00dev->cap_flags))
+ rt2800_bbp_write(rt2x00dev, 4, bbpr4);
+
+ rt2800_bbp_write(rt2x00dev, 21, 0x01);
+ udelay(1);
+ rt2800_bbp_write(rt2x00dev, 21, 0x00);
+
+ rt2800_rf_configrecover(rt2x00dev, rf_store);
+
+ rt2800_register_write(rt2x00dev, TX_PIN_CFG, macorg1);
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, 0x00);
+ rt2800_register_write(rt2x00dev, RF_BYPASS0, 0x00);
+ rt2800_register_write(rt2x00dev, RF_CONTROL0, macorg2);
+ udelay(1);
+ rt2800_register_write(rt2x00dev, RF_BYPASS0, macorg3);
+ rt2800_register_write(rt2x00dev, RF_CONTROL3, macorg4);
+ rt2800_register_write(rt2x00dev, RF_BYPASS3, macorg5);
+ rt2800_register_write(rt2x00dev, MAC_SYS_CTRL, savemacsysctrl);
+ rt2800_register_write(rt2x00dev, 0x13b8, mac13b8);
+}
+
static void rt2800_bbp_core_soft_reset(struct rt2x00_dev *rt2x00dev,
bool set_bw, bool is_ht40)
{
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
@@ -9653,6 +10554,7 @@ static void rt2800_init_rfcsr_6352(struc
rt2800_rxdcoc_calibration(rt2x00dev);
rt2800_bw_filter_calibration(rt2x00dev, true);
rt2800_bw_filter_calibration(rt2x00dev, false);
+ rt2800_loft_iq_calibration(rt2x00dev);
rt2800_rxiq_calibration(rt2x00dev);
}
--- a/drivers/net/wireless/ralink/rt2x00/rt2800lib.h
+++ b/drivers/net/wireless/ralink/rt2x00/rt2800lib.h
@@ -17,6 +17,16 @@
#define WCID_START 33
#define WCID_END 222
#define STA_IDS_SIZE (WCID_END - WCID_START + 2)
+#define CHAIN_0 0x0
+#define CHAIN_1 0x1
+#define RF_ALC_NUM 6
+#define CHAIN_NUM 2
+
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+struct rf_reg_pair {
+ u8 bank;
+ u8 reg;
+ u8 value;
mac80211: rt2x00: experimental improvements for MT7620 wifi Serge Vasilugin reports: To improve mt7620 built-in wifi performance some changes: 1. Correct BW20/BW40 switching (see comments with mark (1)) 2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531 3. Set bbp66 for all chains. 4. US_CYC_CNT init based on Programming guide, default value was 33 (pci), set chipset bus clock with fallback to cpu clock/3. 5. Don't overwrite default values for mt7620. 6. Correct some typos. 7. Add support for external LNA: a) RF and BBP regs never be corrected for this mode b) eLNA is driven the same way as ePA with mt7620's pin PA but vendor driver explicitly pin PA to gpio mode (for forrect calibration?) so I'm not sure that request for pa_pin in dts-file will be enough First 5 changes (really 2) improve performance for boards w/o eLNA/ePA. Changes 7 add support for eLNA Configuration w/o eLAN/ePA and with eLNA show results tx/rx (from router point of view) for each stream: 35-40/30-35 Mbps for HT20 65-70/60-65 Mbps for HT40 Yes. Max results for 2T2R client is 140-145/135-140 with peaks 160/150, It correspond to mediatek driver results. Boards with ePA untested. Reported-by: Serge Vasilugin <vasilugin@yandex.ru> Signed-off-by: Daniel Golle <daniel@makrotopia.org> [directly include v3 of the patchset submitted upstream] (cherry picked from commit 31a6605de04218e1c04bd5c2436c24d7d1c07506) (cherry picked from commit e785ca05e9f0502894772f5df92192b816ba5d7c) (cherry picked from commit 412fcf3d4400f84551f3ead0514834c62d94a251)
2022-07-28 11:40:12 +00:00
+};
/* RT2800 driver data structure */
struct rt2800_drv_data {