[PATCH 0/3] wifi: rtw89: refine SER things of 8851B, 8852C and new state

[PATCH 0/3] wifi: rtw89: refine SER things of 8851B, 8852C and new state

[Ping-Ke Shih]

SER (System Error Recovery) is to detect hardware/firmware abnormal. When
it is detected by firmware, it notifies driver to provide SER IMR and some
registers status to help diagnose the abnormal.

For the best case, we adjust hardware or firmware settings to fix the
problem and avoid this notification. However, some addressed problems
are still needed driver to help recovery; in this case, we don't want
driver show so much information. Patch 3/3 does this thing.

Patch 2/3 is to add pre-/post-M0 states to prevent SER state machine
timeout in driver side while firmware is processing much things.

Chin-Yen Lee (1):
  wifi: rtw89: suppress the log for specific SER called CMDPSR_FRZTO

Zong-Zhe Yang (2):
  wifi: rtw89: pci: fix interrupt enable mask for HALT C2H of RTL8851B
  wifi: rtw89: ser: L1 add pre-M0 and post-M0 states

 drivers/net/wireless/realtek/rtw89/core.h |  2 ++
 drivers/net/wireless/realtek/rtw89/mac.c  | 42 ++++++++++++++++++++++
 drivers/net/wireless/realtek/rtw89/mac.h  |  2 ++
 drivers/net/wireless/realtek/rtw89/pci.c  |  9 +++--
 drivers/net/wireless/realtek/rtw89/pci.h  |  1 +
 drivers/net/wireless/realtek/rtw89/ser.c  | 43 ++++++++++++++++++++++-
 6 files changed, 96 insertions(+), 3 deletions(-)

-- 
2.25.1

[PATCH 1/3] wifi: rtw89: pci: fix interrupt enable mask for HALT C2H of RTL8851B

[Ping-Ke Shih]

From: Zong-Zhe Yang <kevin_yang@realtek.com>

RTL8851B keeps almost the same interrupt flow as RTL8852A and RTL8852B.
But, it uses a different bitmask for interrupt indicator of FW HALT C2H.
So, we make a chip judgement in pci when configuring interrupt mask.

Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
---
 drivers/net/wireless/realtek/rtw89/pci.c | 9 +++++++--
 drivers/net/wireless/realtek/rtw89/pci.h | 1 +
 2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/drivers/net/wireless/realtek/rtw89/pci.c b/drivers/net/wireless/realtek/rtw89/pci.c
index b53f346fef979..92bfef942d3a9 100644
--- a/drivers/net/wireless/realtek/rtw89/pci.c
+++ b/drivers/net/wireless/realtek/rtw89/pci.c
@@ -3216,11 +3216,16 @@ static void rtw89_pci_clear_resource(struct rtw89_dev *rtwdev,
 void rtw89_pci_config_intr_mask(struct rtw89_dev *rtwdev)
 {
 	struct rtw89_pci *rtwpci = (struct rtw89_pci *)rtwdev->priv;
+	const struct rtw89_chip_info *chip = rtwdev->chip;
+	u32 hs0isr_ind_int_en = B_AX_HS0ISR_IND_INT_EN;
+
+	if (chip->chip_id == RTL8851B)
+		hs0isr_ind_int_en = B_AX_HS0ISR_IND_INT_EN_WKARND;
 
 	rtwpci->halt_c2h_intrs = B_AX_HALT_C2H_INT_EN | 0;
 
 	if (rtwpci->under_recovery) {
-		rtwpci->intrs[0] = B_AX_HS0ISR_IND_INT_EN;
+		rtwpci->intrs[0] = hs0isr_ind_int_en;
 		rtwpci->intrs[1] = 0;
 	} else {
 		rtwpci->intrs[0] = B_AX_TXDMA_STUCK_INT_EN |
@@ -3230,7 +3235,7 @@ void rtw89_pci_config_intr_mask(struct rtw89_dev *rtwdev)
 				   B_AX_RXDMA_STUCK_INT_EN |
 				   B_AX_RDU_INT_EN |
 				   B_AX_RPQBD_FULL_INT_EN |
-				   B_AX_HS0ISR_IND_INT_EN;
+				   hs0isr_ind_int_en;
 
 		rtwpci->intrs[1] = B_AX_HC10ISR_IND_INT_EN;
 	}
diff --git a/drivers/net/wireless/realtek/rtw89/pci.h b/drivers/net/wireless/realtek/rtw89/pci.h
index 0e4bd210b100f..2f3d1ad3b0f7d 100644
--- a/drivers/net/wireless/realtek/rtw89/pci.h
+++ b/drivers/net/wireless/realtek/rtw89/pci.h
@@ -150,6 +150,7 @@
 #define B_AX_HD1ISR_IND_INT_EN		BIT(26)
 #define B_AX_HD0ISR_IND_INT_EN		BIT(25)
 #define B_AX_HS0ISR_IND_INT_EN		BIT(24)
+#define B_AX_HS0ISR_IND_INT_EN_WKARND	BIT(23)
 #define B_AX_RETRAIN_INT_EN		BIT(21)
 #define B_AX_RPQBD_FULL_INT_EN		BIT(20)
 #define B_AX_RDU_INT_EN			BIT(19)
-- 
2.25.1

[PATCH 2/3] wifi: rtw89: ser: L1 add pre-M0 and post-M0 states

[Ping-Ke Shih]

From: Zong-Zhe Yang <kevin_yang@realtek.com>

Newer FW re-design SER (syetem error recovery) L1 (level 1) flow.
New L1 flow will expect two extra states before original L1 flow.

* Before:
fw --- M1 --> driver
fw <-- M2 --- driver
fw --- M3 --> driver
fw <-- M4 --- driver
fw --- M5 --> driver

* After:
fw --- pre-M0  --> driver
fw <-- post-M0 --- driver
fw --- M1 --> driver
fw <-- M2 --- driver
fw --- M3 --> driver
fw <-- M4 --- driver
fw --- M5 --> driver

Then before M1, FW gets one more interval to deal with things that FW
should have handled well. To consider backward/forward compatibility,
FW and driver won't change flow from M1 to M5. (only except that halt
trigger control will change a little bit.) So, there will be two differnt
starting points of SER L1.
* old FW: SER L1 starts from M1
* new FW: SER L1 starts from pre-M0
Then, driver adds the new SER L1 entry and also keep the original one
instead of changing it.

Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
---
 drivers/net/wireless/realtek/rtw89/core.h |  1 +
 drivers/net/wireless/realtek/rtw89/mac.c  |  6 ++++
 drivers/net/wireless/realtek/rtw89/mac.h  |  2 ++
 drivers/net/wireless/realtek/rtw89/ser.c  | 43 ++++++++++++++++++++++-
 4 files changed, 51 insertions(+), 1 deletion(-)

diff --git a/drivers/net/wireless/realtek/rtw89/core.h b/drivers/net/wireless/realtek/rtw89/core.h
index f697902706acc..b71827751848a 100644
--- a/drivers/net/wireless/realtek/rtw89/core.h
+++ b/drivers/net/wireless/realtek/rtw89/core.h
@@ -3945,6 +3945,7 @@ enum rtw89_ser_rcvy_step {
 struct rtw89_ser {
 	u8 state;
 	u8 alarm_event;
+	bool prehandle_l1;
 
 	struct work_struct ser_hdl_work;
 	struct delayed_work ser_alarm_work;
diff --git a/drivers/net/wireless/realtek/rtw89/mac.c b/drivers/net/wireless/realtek/rtw89/mac.c
index 64dc36470840c..e5996a94c0def 100644
--- a/drivers/net/wireless/realtek/rtw89/mac.c
+++ b/drivers/net/wireless/realtek/rtw89/mac.c
@@ -676,6 +676,7 @@ EXPORT_SYMBOL(rtw89_mac_get_err_status);
 
 int rtw89_mac_set_err_status(struct rtw89_dev *rtwdev, u32 err)
 {
+	struct rtw89_ser *ser = &rtwdev->ser;
 	u32 halt;
 	int ret = 0;
 
@@ -692,6 +693,11 @@ int rtw89_mac_set_err_status(struct rtw89_dev *rtwdev, u32 err)
 	}
 
 	rtw89_write32(rtwdev, R_AX_HALT_H2C, err);
+
+	if (ser->prehandle_l1 &&
+	    (err == MAC_AX_ERR_L1_DISABLE_EN || err == MAC_AX_ERR_L1_RCVY_EN))
+		return 0;
+
 	rtw89_write32(rtwdev, R_AX_HALT_H2C_CTRL, B_AX_HALT_H2C_TRIGGER);
 
 	return 0;
diff --git a/drivers/net/wireless/realtek/rtw89/mac.h b/drivers/net/wireless/realtek/rtw89/mac.h
index d3922d4fe288a..e340720bbb889 100644
--- a/drivers/net/wireless/realtek/rtw89/mac.h
+++ b/drivers/net/wireless/realtek/rtw89/mac.h
@@ -642,6 +642,7 @@ enum mac_ax_err_info {
 	MAC_AX_ERR_L0_PROMOTE_TO_L1 = 0x0010,
 
 	/* L1 */
+	MAC_AX_ERR_L1_PREERR_DMAC = 0x999,
 	MAC_AX_ERR_L1_ERR_DMAC = 0x1000,
 	MAC_AX_ERR_L1_RESET_DISABLE_DMAC_DONE = 0x1001,
 	MAC_AX_ERR_L1_RESET_RECOVERY_DONE = 0x1002,
@@ -780,6 +781,7 @@ enum mac_ax_err_info {
 	MAC_AX_ERR_L1_RCVY_EN = 0x0002,
 	MAC_AX_ERR_L1_RCVY_STOP_REQ = 0x0003,
 	MAC_AX_ERR_L1_RCVY_START_REQ = 0x0004,
+	MAC_AX_ERR_L1_RESET_START_DMAC = 0x000A,
 	MAC_AX_ERR_L0_CFG_NOTIFY = 0x0010,
 	MAC_AX_ERR_L0_CFG_DIS_NOTIFY = 0x0011,
 	MAC_AX_ERR_L0_CFG_HANDSHAKE = 0x0012,
diff --git a/drivers/net/wireless/realtek/rtw89/ser.c b/drivers/net/wireless/realtek/rtw89/ser.c
index 9e9f6947e7f14..9ba99f3764e7e 100644
--- a/drivers/net/wireless/realtek/rtw89/ser.c
+++ b/drivers/net/wireless/realtek/rtw89/ser.c
@@ -20,12 +20,14 @@ enum ser_evt {
 	SER_EV_NONE,
 	SER_EV_STATE_IN,
 	SER_EV_STATE_OUT,
+	SER_EV_L1_RESET_PREPARE, /* pre-M0 */
 	SER_EV_L1_RESET, /* M1 */
 	SER_EV_DO_RECOVERY, /* M3 */
 	SER_EV_MAC_RESET_DONE, /* M5 */
 	SER_EV_L2_RESET,
 	SER_EV_L2_RECFG_DONE,
 	SER_EV_L2_RECFG_TIMEOUT,
+	SER_EV_M1_TIMEOUT,
 	SER_EV_M3_TIMEOUT,
 	SER_EV_FW_M5_TIMEOUT,
 	SER_EV_L0_RESET,
@@ -34,6 +36,7 @@ enum ser_evt {
 
 enum ser_state {
 	SER_IDLE_ST,
+	SER_L1_RESET_PRE_ST,
 	SER_RESET_TRX_ST,
 	SER_DO_HCI_ST,
 	SER_L2_RESET_ST,
@@ -374,6 +377,13 @@ static int hal_stop_dma(struct rtw89_ser *ser)
 	return ret;
 }
 
+static void hal_send_post_m0_event(struct rtw89_ser *ser)
+{
+	struct rtw89_dev *rtwdev = container_of(ser, struct rtw89_dev, ser);
+
+	rtw89_mac_set_err_status(rtwdev, MAC_AX_ERR_L1_RESET_START_DMAC);
+}
+
 static void hal_send_m2_event(struct rtw89_ser *ser)
 {
 	struct rtw89_dev *rtwdev = container_of(ser, struct rtw89_dev, ser);
@@ -398,6 +408,9 @@ static void ser_idle_st_hdl(struct rtw89_ser *ser, u8 evt)
 		rtw89_hci_recovery_complete(rtwdev);
 		clear_bit(RTW89_FLAG_CRASH_SIMULATING, rtwdev->flags);
 		break;
+	case SER_EV_L1_RESET_PREPARE:
+		ser_state_goto(ser, SER_L1_RESET_PRE_ST);
+		break;
 	case SER_EV_L1_RESET:
 		ser_state_goto(ser, SER_RESET_TRX_ST);
 		break;
@@ -412,6 +425,28 @@ static void ser_idle_st_hdl(struct rtw89_ser *ser, u8 evt)
 	}
 }
 
+static void ser_l1_reset_pre_st_hdl(struct rtw89_ser *ser, u8 evt)
+{
+	switch (evt) {
+	case SER_EV_STATE_IN:
+		ser->prehandle_l1 = true;
+		hal_send_post_m0_event(ser);
+		ser_set_alarm(ser, 1000, SER_EV_M1_TIMEOUT);
+		break;
+	case SER_EV_L1_RESET:
+		ser_state_goto(ser, SER_RESET_TRX_ST);
+		break;
+	case SER_EV_M1_TIMEOUT:
+		ser_state_goto(ser, SER_L2_RESET_ST);
+		break;
+	case SER_EV_STATE_OUT:
+		ser_del_alarm(ser);
+		break;
+	default:
+		break;
+	}
+}
+
 static void ser_reset_trx_st_hdl(struct rtw89_ser *ser, u8 evt)
 {
 	struct rtw89_dev *rtwdev = container_of(ser, struct rtw89_dev, ser);
@@ -654,12 +689,14 @@ static const struct event_ent ser_ev_tbl[] = {
 	{SER_EV_NONE, "SER_EV_NONE"},
 	{SER_EV_STATE_IN, "SER_EV_STATE_IN"},
 	{SER_EV_STATE_OUT, "SER_EV_STATE_OUT"},
-	{SER_EV_L1_RESET, "SER_EV_L1_RESET"},
+	{SER_EV_L1_RESET_PREPARE, "SER_EV_L1_RESET_PREPARE pre-m0"},
+	{SER_EV_L1_RESET, "SER_EV_L1_RESET m1"},
 	{SER_EV_DO_RECOVERY, "SER_EV_DO_RECOVERY m3"},
 	{SER_EV_MAC_RESET_DONE, "SER_EV_MAC_RESET_DONE m5"},
 	{SER_EV_L2_RESET, "SER_EV_L2_RESET"},
 	{SER_EV_L2_RECFG_DONE, "SER_EV_L2_RECFG_DONE"},
 	{SER_EV_L2_RECFG_TIMEOUT, "SER_EV_L2_RECFG_TIMEOUT"},
+	{SER_EV_M1_TIMEOUT, "SER_EV_M1_TIMEOUT"},
 	{SER_EV_M3_TIMEOUT, "SER_EV_M3_TIMEOUT"},
 	{SER_EV_FW_M5_TIMEOUT, "SER_EV_FW_M5_TIMEOUT"},
 	{SER_EV_L0_RESET, "SER_EV_L0_RESET"},
@@ -668,6 +705,7 @@ static const struct event_ent ser_ev_tbl[] = {
 
 static const struct state_ent ser_st_tbl[] = {
 	{SER_IDLE_ST, "SER_IDLE_ST", ser_idle_st_hdl},
+	{SER_L1_RESET_PRE_ST, "SER_L1_RESET_PRE_ST", ser_l1_reset_pre_st_hdl},
 	{SER_RESET_TRX_ST, "SER_RESET_TRX_ST", ser_reset_trx_st_hdl},
 	{SER_DO_HCI_ST, "SER_DO_HCI_ST", ser_do_hci_st_hdl},
 	{SER_L2_RESET_ST, "SER_L2_RESET_ST", ser_l2_reset_st_hdl}
@@ -713,6 +751,9 @@ int rtw89_ser_notify(struct rtw89_dev *rtwdev, u32 err)
 	rtw89_info(rtwdev, "SER catches error: 0x%x\n", err);
 
 	switch (err) {
+	case MAC_AX_ERR_L1_PREERR_DMAC: /* pre-M0 */
+		event = SER_EV_L1_RESET_PREPARE;
+		break;
 	case MAC_AX_ERR_L1_ERR_DMAC:
 	case MAC_AX_ERR_L0_PROMOTE_TO_L1:
 		event = SER_EV_L1_RESET; /* M1 */
-- 
2.25.1

[PATCH 3/3] wifi: rtw89: suppress the log for specific SER called CMDPSR_FRZTO

[Ping-Ke Shih]

From: Chin-Yen Lee <timlee@realtek.com>

For 8852CE, there is abnormal state called CMDPSR_FRZTO,
which occasionally happens in some platforms, and could be
found by firmware and fixed in current SER flow, so we add
suppress function to avoid verbose message for this resolved case.

Signed-off-by: Chin-Yen Lee <timlee@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
---
 drivers/net/wireless/realtek/rtw89/core.h |  1 +
 drivers/net/wireless/realtek/rtw89/mac.c  | 36 +++++++++++++++++++++++
 2 files changed, 37 insertions(+)

diff --git a/drivers/net/wireless/realtek/rtw89/core.h b/drivers/net/wireless/realtek/rtw89/core.h
index b71827751848a..154fec1a231d1 100644
--- a/drivers/net/wireless/realtek/rtw89/core.h
+++ b/drivers/net/wireless/realtek/rtw89/core.h
@@ -3939,6 +3939,7 @@ enum rtw89_ser_rcvy_step {
 	RTW89_SER_DRV_STOP_RX,
 	RTW89_SER_DRV_STOP_RUN,
 	RTW89_SER_HAL_STOP_DMA,
+	RTW89_SER_SUPPRESS_LOG,
 	RTW89_NUM_OF_SER_FLAGS
 };
 
diff --git a/drivers/net/wireless/realtek/rtw89/mac.c b/drivers/net/wireless/realtek/rtw89/mac.c
index e5996a94c0def..acba53cf7cc31 100644
--- a/drivers/net/wireless/realtek/rtw89/mac.c
+++ b/drivers/net/wireless/realtek/rtw89/mac.c
@@ -644,6 +644,39 @@ static void rtw89_mac_dump_err_status(struct rtw89_dev *rtwdev,
 	rtw89_info(rtwdev, "<---\n");
 }
 
+static bool rtw89_mac_suppress_log(struct rtw89_dev *rtwdev, u32 err)
+{
+	struct rtw89_ser *ser = &rtwdev->ser;
+	u32 dmac_err, imr, isr;
+	int ret;
+
+	if (rtwdev->chip->chip_id == RTL8852C) {
+		ret = rtw89_mac_check_mac_en(rtwdev, 0, RTW89_DMAC_SEL);
+		if (ret)
+			return true;
+
+		if (err == MAC_AX_ERR_L1_ERR_DMAC) {
+			dmac_err = rtw89_read32(rtwdev, R_AX_DMAC_ERR_ISR);
+			imr = rtw89_read32(rtwdev, R_AX_TXPKTCTL_B0_ERRFLAG_IMR);
+			isr = rtw89_read32(rtwdev, R_AX_TXPKTCTL_B0_ERRFLAG_ISR);
+
+			if ((dmac_err & B_AX_TXPKTCTRL_ERR_FLAG) &&
+			    ((isr & imr) & B_AX_B0_ISR_ERR_CMDPSR_FRZTO)) {
+				set_bit(RTW89_SER_SUPPRESS_LOG, ser->flags);
+				return true;
+			}
+		} else if (err == MAC_AX_ERR_L1_RESET_DISABLE_DMAC_DONE) {
+			if (test_bit(RTW89_SER_SUPPRESS_LOG, ser->flags))
+				return true;
+		} else if (err == MAC_AX_ERR_L1_RESET_RECOVERY_DONE) {
+			if (test_and_clear_bit(RTW89_SER_SUPPRESS_LOG, ser->flags))
+				return true;
+		}
+	}
+
+	return false;
+}
+
 u32 rtw89_mac_get_err_status(struct rtw89_dev *rtwdev)
 {
 	u32 err, err_scnr;
@@ -667,6 +700,9 @@ u32 rtw89_mac_get_err_status(struct rtw89_dev *rtwdev)
 	else if (err_scnr == RTW89_RXI300_ERROR)
 		err = MAC_AX_ERR_RXI300;
 
+	if (rtw89_mac_suppress_log(rtwdev, err))
+		return err;
+
 	rtw89_fw_st_dbg_dump(rtwdev);
 	rtw89_mac_dump_err_status(rtwdev, err);
 
-- 
2.25.1

Re: [PATCH 1/3] wifi: rtw89: pci: fix interrupt enable mask for HALT C2H of RTL8851B

[Kalle Valo]

Ping-Ke Shih <pkshih@realtek.com> wrote:

> From: Zong-Zhe Yang <kevin_yang@realtek.com>
> 
> RTL8851B keeps almost the same interrupt flow as RTL8852A and RTL8852B.
> But, it uses a different bitmask for interrupt indicator of FW HALT C2H.
> So, we make a chip judgement in pci when configuring interrupt mask.
> 
> Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
> Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>

3 patches applied to wireless-next.git, thanks.

aa70fa4f7dd8 wifi: rtw89: pci: fix interrupt enable mask for HALT C2H of RTL8851B
56617fd02adb wifi: rtw89: ser: L1 add pre-M0 and post-M0 states
8130e94e888b wifi: rtw89: suppress the log for specific SER called CMDPSR_FRZTO

-- 
https://patchwork.kernel.org/project/linux-wireless/patch/20230508084335.42953-2-pkshih@realtek.com/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches