[PATCH 0/6] ath9k patches

[PATCH 0/6] ath9k patches

[Sujith Manoharan]

From: Sujith Manoharan <c_manoha@qca.qualcomm.com>

ath9k patches for -next.

Sujith Manoharan (6):
  ath9k: Update QCA953x initvals
  ath9k: Update AR955x initvals
  ath9k: Add a macro to identify PCOEM chips
  ath9k: Fix manual peak calibration initialization
  ath9k: Set correct peak detect threshold
  ath9k: Enable manual peak detect calibration

 drivers/net/wireless/ath/ath9k/ar9003_calib.c      | 61 +++++++++++++++++-----
 drivers/net/wireless/ath/ath9k/ar953x_initvals.h   |  4 +-
 .../net/wireless/ath/ath9k/ar955x_1p0_initvals.h   |  4 +-
 drivers/net/wireless/ath/ath9k/reg.h               |  3 ++
 4 files changed, 55 insertions(+), 17 deletions(-)

-- 
2.2.2

[PATCH 1/6] ath9k: Update QCA953x initvals

[Sujith Manoharan]

From: Sujith Manoharan <c_manoha@qca.qualcomm.com>

This patch disables HW peak detect calibration for
QCA953x, since it is problematic on a few boards.
Instead, manual calibration will be done in the
driver.

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
---
 drivers/net/wireless/ath/ath9k/ar953x_initvals.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/wireless/ath/ath9k/ar953x_initvals.h b/drivers/net/wireless/ath/ath9k/ar953x_initvals.h
index 159cc6f..6fc0d07 100644
--- a/drivers/net/wireless/ath/ath9k/ar953x_initvals.h
+++ b/drivers/net/wireless/ath/ath9k/ar953x_initvals.h
@@ -358,7 +358,7 @@ static const u32 qca953x_1p0_baseband_postamble[][5] = {
 	{0x00009e1c, 0x0001cf9c, 0x0001cf9c, 0x00021f9c, 0x00021f9c},
 	{0x00009e20, 0x000003b5, 0x000003b5, 0x000003ce, 0x000003ce},
 	{0x00009e2c, 0x0000001c, 0x0000001c, 0x00000021, 0x00000021},
-	{0x00009e3c, 0xcfa10820, 0xcfa10820, 0xcfa10822, 0xcfa10822},
+	{0x00009e3c, 0xcfa10820, 0xcfa10820, 0xcfa10820, 0xcfa10820},
 	{0x00009e44, 0xfe321e27, 0xfe321e27, 0xfe291e27, 0xfe291e27},
 	{0x00009e48, 0x5030201a, 0x5030201a, 0x50302012, 0x50302012},
 	{0x00009fc8, 0x0003f000, 0x0003f000, 0x0001a000, 0x0001a000},
@@ -378,7 +378,7 @@ static const u32 qca953x_1p0_baseband_postamble[][5] = {
 	{0x0000a284, 0x00000000, 0x00000000, 0x00000010, 0x00000010},
 	{0x0000a288, 0x00000110, 0x00000110, 0x00000110, 0x00000110},
 	{0x0000a28c, 0x00022222, 0x00022222, 0x00022222, 0x00022222},
-	{0x0000a2c4, 0x00158d18, 0x00158d18, 0x00158d18, 0x00158d18},
+	{0x0000a2c4, 0x00158d18, 0x00158d18, 0x00058d18, 0x00058d18},
 	{0x0000a2cc, 0x18c50033, 0x18c43433, 0x18c41033, 0x18c44c33},
 	{0x0000a2d0, 0x00041982, 0x00041982, 0x00041982, 0x00041982},
 	{0x0000a2d8, 0x7999a83b, 0x7999a83b, 0x7999a83b, 0x7999a83b},
-- 
2.2.2

[PATCH 2/6] ath9k: Update AR955x initvals

[Sujith Manoharan]

From: Sujith Manoharan <c_manoha@qca.qualcomm.com>

This patch disables HW peak detect calibration for
AR955x, since it is problematic on a few boards.
Instead, manual calibration will be done in the
driver.

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
---
 drivers/net/wireless/ath/ath9k/ar955x_1p0_initvals.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/wireless/ath/ath9k/ar955x_1p0_initvals.h b/drivers/net/wireless/ath/ath9k/ar955x_1p0_initvals.h
index fd6a84c..148562a 100644
--- a/drivers/net/wireless/ath/ath9k/ar955x_1p0_initvals.h
+++ b/drivers/net/wireless/ath/ath9k/ar955x_1p0_initvals.h
@@ -63,7 +63,7 @@ static const u32 ar955x_1p0_baseband_postamble[][5] = {
 	{0x00009e1c, 0x0001cf9c, 0x0001cf9c, 0x00021f9c, 0x00021f9c},
 	{0x00009e20, 0x000003b5, 0x000003b5, 0x000003ce, 0x000003ce},
 	{0x00009e2c, 0x0000001c, 0x0000001c, 0x00000021, 0x00000021},
-	{0x00009e3c, 0xcfa10820, 0xcfa10820, 0xcfa10822, 0xcfa10822},
+	{0x00009e3c, 0xcfa10820, 0xcfa10820, 0xcfa10820, 0xcfa10820},
 	{0x00009e44, 0xfe321e27, 0xfe321e27, 0xfe291e27, 0xfe291e27},
 	{0x00009e48, 0x5030201a, 0x5030201a, 0x50302012, 0x50302012},
 	{0x00009fc8, 0x0003f000, 0x0003f000, 0x0001a000, 0x0001a000},
@@ -83,7 +83,7 @@ static const u32 ar955x_1p0_baseband_postamble[][5] = {
 	{0x0000a284, 0x00000000, 0x00000000, 0x00000010, 0x00000010},
 	{0x0000a288, 0x00000110, 0x00000110, 0x00000110, 0x00000110},
 	{0x0000a28c, 0x00022222, 0x00022222, 0x00022222, 0x00022222},
-	{0x0000a2c4, 0x00158d18, 0x00158d18, 0x00158d18, 0x00158d18},
+	{0x0000a2c4, 0x00158d18, 0x00158d18, 0x00058d18, 0x00058d18},
 	{0x0000a2cc, 0x18c50033, 0x18c43433, 0x18c41033, 0x18c44c33},
 	{0x0000a2d0, 0x00041982, 0x00041982, 0x00041982, 0x00041982},
 	{0x0000a2d8, 0x7999a83b, 0x7999a83b, 0x7999a83b, 0x7999a83b},
-- 
2.2.2

[PATCH 3/6] ath9k: Add a macro to identify PCOEM chips

[Sujith Manoharan]

From: Sujith Manoharan <c_manoha@qca.qualcomm.com>

This can be use if we need to apply register settings
for all PCOEM solutions (in the AR9003 family).

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
---
 drivers/net/wireless/ath/ath9k/reg.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/net/wireless/ath/ath9k/reg.h b/drivers/net/wireless/ath/ath9k/reg.h
index eb2bb0d..441145e 100644
--- a/drivers/net/wireless/ath/ath9k/reg.h
+++ b/drivers/net/wireless/ath/ath9k/reg.h
@@ -900,10 +900,13 @@
 	(((_ah)->hw_version.macVersion == AR_SREV_VERSION_9485))
 #define AR_SREV_9565(_ah) \
 	(((_ah)->hw_version.macVersion == AR_SREV_VERSION_9565))
+#define AR_SREV_9003_PCOEM(_ah) \
+	(AR_SREV_9462(_ah) || AR_SREV_9485(_ah) || AR_SREV_9565(_ah))
 #else
 #define AR_SREV_9462(_ah) 0
 #define AR_SREV_9485(_ah) 0
 #define AR_SREV_9565(_ah) 0
+#define AR_SREV_9003_PCOEM(_ah) )
 #endif
 
 #define AR_SREV_9485_11_OR_LATER(_ah) \
-- 
2.2.2

[PATCH 4/6] ath9k: Fix manual peak calibration initialization

[Sujith Manoharan]

From: Sujith Manoharan <c_manoha@qca.qualcomm.com>

The LNA gain setting override needs to be done
only for AR9330 and PCOEM chips.

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
---
 drivers/net/wireless/ath/ath9k/ar9003_calib.c | 38 +++++++++++++++++++++------
 1 file changed, 30 insertions(+), 8 deletions(-)

diff --git a/drivers/net/wireless/ath/ath9k/ar9003_calib.c b/drivers/net/wireless/ath/ath9k/ar9003_calib.c
index 06ab71d..374aef1 100644
--- a/drivers/net/wireless/ath/ath9k/ar9003_calib.c
+++ b/drivers/net/wireless/ath/ath9k/ar9003_calib.c
@@ -1205,22 +1205,34 @@ static void ar9003_hw_manual_peak_cal(struct ath_hw *ah, u8 chain, bool is_2g)
 	int offset[8] = {0}, total = 0, test;
 	int agc_out, i;
 
+	/*
+	 * Turn off LNA/SW.
+	 */
 	REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_GAINSTAGES(chain),
 		      AR_PHY_65NM_RXRF_GAINSTAGES_RX_OVERRIDE, 0x1);
 	REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_GAINSTAGES(chain),
 		      AR_PHY_65NM_RXRF_GAINSTAGES_LNAON_CALDC, 0x0);
-	if (is_2g)
-		REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_GAINSTAGES(chain),
-			      AR_PHY_65NM_RXRF_GAINSTAGES_LNA2G_GAIN_OVR, 0x0);
-	else
-		REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_GAINSTAGES(chain),
-			      AR_PHY_65NM_RXRF_GAINSTAGES_LNA5G_GAIN_OVR, 0x0);
 
+	if (AR_SREV_9003_PCOEM(ah) || AR_SREV_9330_11(ah)) {
+		if (is_2g)
+			REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_GAINSTAGES(chain),
+				      AR_PHY_65NM_RXRF_GAINSTAGES_LNA2G_GAIN_OVR, 0x0);
+		else
+			REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_GAINSTAGES(chain),
+				      AR_PHY_65NM_RXRF_GAINSTAGES_LNA5G_GAIN_OVR, 0x0);
+	}
+
+	/*
+	 * Turn off RXON.
+	 */
 	REG_RMW_FIELD(ah, AR_PHY_65NM_RXTX2(chain),
 		      AR_PHY_65NM_RXTX2_RXON_OVR, 0x1);
 	REG_RMW_FIELD(ah, AR_PHY_65NM_RXTX2(chain),
 		      AR_PHY_65NM_RXTX2_RXON, 0x0);
 
+	/*
+	 * Turn on AGC for cal.
+	 */
 	REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_AGC(chain),
 		      AR_PHY_65NM_RXRF_AGC_AGC_OVERRIDE, 0x1);
 	REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_AGC(chain),
@@ -1228,10 +1240,11 @@ static void ar9003_hw_manual_peak_cal(struct ath_hw *ah, u8 chain, bool is_2g)
 	REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_AGC(chain),
 		      AR_PHY_65NM_RXRF_AGC_AGC_CAL_OVR, 0x1);
 
-	if (AR_SREV_9330_11(ah)) {
+	if (AR_SREV_9330_11(ah))
 		REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_AGC(chain),
 			      AR_PHY_65NM_RXRF_AGC_AGC2G_CALDAC_OVR, 0x0);
-	} else {
+
+	if (AR_SREV_9003_PCOEM(ah)) {
 		if (is_2g)
 			REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_AGC(chain),
 				      AR_PHY_65NM_RXRF_AGC_AGC2G_DBDAC_OVR, 0x0);
@@ -1266,10 +1279,19 @@ static void ar9003_hw_manual_peak_cal(struct ath_hw *ah, u8 chain, bool is_2g)
 		REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_AGC(chain),
 			      AR_PHY_65NM_RXRF_AGC_AGC5G_CALDAC_OVR, total);
 
+	/*
+	 * Turn on LNA.
+	 */
 	REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_GAINSTAGES(chain),
 		      AR_PHY_65NM_RXRF_GAINSTAGES_RX_OVERRIDE, 0);
+	/*
+	 * Turn off RXON.
+	 */
 	REG_RMW_FIELD(ah, AR_PHY_65NM_RXTX2(chain),
 		      AR_PHY_65NM_RXTX2_RXON_OVR, 0);
+	/*
+	 * Turn off peak detect calibration.
+	 */
 	REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_AGC(chain),
 		      AR_PHY_65NM_RXRF_AGC_AGC_CAL_OVR, 0);
 }
-- 
2.2.2

[PATCH 5/6] ath9k: Set correct peak detect threshold

[Sujith Manoharan]

From: Sujith Manoharan <c_manoha@qca.qualcomm.com>

The value is different for PCOEM cards and AR955x/AR953x.

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
---
 drivers/net/wireless/ath/ath9k/ar9003_calib.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

diff --git a/drivers/net/wireless/ath/ath9k/ar9003_calib.c b/drivers/net/wireless/ath/ath9k/ar9003_calib.c
index 374aef1..91130c0 100644
--- a/drivers/net/wireless/ath/ath9k/ar9003_calib.c
+++ b/drivers/net/wireless/ath/ath9k/ar9003_calib.c
@@ -1203,7 +1203,12 @@ static void ar9003_hw_tx_iq_cal_reload(struct ath_hw *ah)
 static void ar9003_hw_manual_peak_cal(struct ath_hw *ah, u8 chain, bool is_2g)
 {
 	int offset[8] = {0}, total = 0, test;
-	int agc_out, i;
+	int agc_out, i, peak_detect_threshold;
+
+	if (AR_SREV_9550(ah) || AR_SREV_9531(ah))
+		peak_detect_threshold = 8;
+	else
+		peak_detect_threshold = 0;
 
 	/*
 	 * Turn off LNA/SW.
@@ -1244,13 +1249,15 @@ static void ar9003_hw_manual_peak_cal(struct ath_hw *ah, u8 chain, bool is_2g)
 		REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_AGC(chain),
 			      AR_PHY_65NM_RXRF_AGC_AGC2G_CALDAC_OVR, 0x0);
 
-	if (AR_SREV_9003_PCOEM(ah)) {
+	if (AR_SREV_9003_PCOEM(ah) || AR_SREV_9550(ah) || AR_SREV_9531(ah)) {
 		if (is_2g)
 			REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_AGC(chain),
-				      AR_PHY_65NM_RXRF_AGC_AGC2G_DBDAC_OVR, 0x0);
+				      AR_PHY_65NM_RXRF_AGC_AGC2G_DBDAC_OVR,
+				      peak_detect_threshold);
 		else
 			REG_RMW_FIELD(ah, AR_PHY_65NM_RXRF_AGC(chain),
-				      AR_PHY_65NM_RXRF_AGC_AGC5G_DBDAC_OVR, 0x0);
+				      AR_PHY_65NM_RXRF_AGC_AGC5G_DBDAC_OVR,
+				      peak_detect_threshold);
 	}
 
 	for (i = 6; i > 0; i--) {
-- 
2.2.2

[PATCH 6/6] ath9k: Enable manual peak detect calibration

[Sujith Manoharan]

From: Sujith Manoharan <c_manoha@qca.qualcomm.com>

On some AR955x/QCA953x boards, noise floor calibration
gets stuck and the cause is a hardware/BB issue. To fix
this, peak detect calibration in the HW is disabled
and is done in the driver instead. There a few differences
with the calibration routine for older chips like
AR9331.

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
---
 drivers/net/wireless/ath/ath9k/ar9003_calib.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/drivers/net/wireless/ath/ath9k/ar9003_calib.c b/drivers/net/wireless/ath/ath9k/ar9003_calib.c
index 91130c0..174442b 100644
--- a/drivers/net/wireless/ath/ath9k/ar9003_calib.c
+++ b/drivers/net/wireless/ath/ath9k/ar9003_calib.c
@@ -1640,8 +1640,14 @@ static bool ar9003_hw_init_cal_soc(struct ath_hw *ah,
 
 skip_tx_iqcal:
 	if (run_agc_cal || !(ah->ah_flags & AH_FASTCC)) {
-		if (AR_SREV_9330_11(ah))
-			ar9003_hw_manual_peak_cal(ah, 0, IS_CHAN_2GHZ(chan));
+		if (AR_SREV_9330_11(ah) || AR_SREV_9531(ah) || AR_SREV_9550(ah)) {
+			for (i = 0; i < AR9300_MAX_CHAINS; i++) {
+				if (!(ah->rxchainmask & (1 << i)))
+					continue;
+				ar9003_hw_manual_peak_cal(ah, i,
+							  IS_CHAN_2GHZ(chan));
+			}
+		}
 
 		/*
 		 * For non-AR9550 chips, we just trigger AGC calibration
-- 
2.2.2

Re: [PATCH 3/6] ath9k: Add a macro to identify PCOEM chips

[Sujith Manoharan]

Sujith Manoharan wrote:
>  #define AR_SREV_9462(_ah) 0
>  #define AR_SREV_9485(_ah) 0
>  #define AR_SREV_9565(_ah) 0
> +#define AR_SREV_9003_PCOEM(_ah) )

Kalle, this is a typo - I'll resend the series.

Sujith

Re: [PATCH 3/6] ath9k: Add a macro to identify PCOEM chips

[Kalle Valo]

Sujith Manoharan <sujith@msujith.org> writes:

> Sujith Manoharan wrote:
>>  #define AR_SREV_9462(_ah) 0
>>  #define AR_SREV_9485(_ah) 0
>>  #define AR_SREV_9565(_ah) 0
>> +#define AR_SREV_9003_PCOEM(_ah) )
>
> Kalle, this is a typo - I'll resend the series.

Ok. I had actually put these into "Not Applicable" by accident and
that's why I had missed them.

-- 
Kalle Valo

Re: [PATCH 3/6] ath9k: Add a macro to identify PCOEM chips

[Sujith Manoharan]

Kalle Valo wrote:
> > Kalle, this is a typo - I'll resend the series.
> 
> Ok. I had actually put these into "Not Applicable" by accident and
> that's why I had missed them.

Ah ok. I thought I had to enable some patchwork filters manually to see
the pending ath9k patches.

Sujith

Re: [PATCH 3/6] ath9k: Add a macro to identify PCOEM chips

[Kalle Valo]

Sujith Manoharan <sujith@msujith.org> writes:

> Kalle Valo wrote:
>> > Kalle, this is a typo - I'll resend the series.
>> 
>> Ok. I had actually put these into "Not Applicable" by accident and
>> that's why I had missed them.
>
> Ah ok. I thought I had to enable some patchwork filters manually to see
> the pending ath9k patches.

For ath10k there's currently a separate patchwork project (I haven't
bothered to change that yet), but for all other drivers you should see
the patches from this link:

https://patchwork.kernel.org/project/linux-wireless/list/

If the patch is not in the list, it's not commited and there's no clear
reason why it was dropped, the chances are that I did a mistake. Like in
this case and also with one of the mwifiex patch earlier this week. I'll
need to be more careful with the states.

-- 
Kalle Valo

Re: [PATCH 3/6] ath9k: Add a macro to identify PCOEM chips

[Sujith Manoharan]

Kalle Valo wrote:
> For ath10k there's currently a separate patchwork project (I haven't
> bothered to change that yet), but for all other drivers you should see
> the patches from this link:
> 
> https://patchwork.kernel.org/project/linux-wireless/list/
> 
> If the patch is not in the list, it's not commited and there's no clear
> reason why it was dropped, the chances are that I did a mistake. Like in
> this case and also with one of the mwifiex patch earlier this week. I'll
> need to be more careful with the states.

Ok, thanks.

Sujith