After this patchset, basic functions are ready, so we can enable 8852ce.
It can play as STA, AP and monitor modes. The BT coexistence and power save
are still cooking, but these don't affect the performance of Wi-Fi only.
These patchset contains a patch to send RF calibration settings to
firmware. Patches 2-3 are to add necessary chip_ops and chip_info.
Patch 4 is to fine tune polling interval to enter/leave low power mode.
Since 8852ce will do this more frequently, polling interval become more
important.
Patch 5-6 are to correct registers settings found when we develop 8852ce.
Patch 7 is to fix compiler warnings reported by clang.
Final patch is to add 8852ce to Makefile and Kconfig, and then user can
use it with firmware version 0.27.20.0 that has been merged.
Ping-Ke Shih (8):
rtw89: 8852c: rfk: get calibrated channels to notify firmware
rtw89: 8852c: add chip_ops::bb_ctrl_btc_preagc
rtw89: 8852c: add basic and remaining chip_info
rtw89: ps: fine tune polling interval while changing low power mode
rtw89: correct AID settings of beamformee
rtw89: 8852c: correct register definitions used by 8852c
rtw89: 8852c: fix warning of FIELD_PREP() mask type
rtw89: 8852c: add 8852ce to Makefile and Kconfig
drivers/net/wireless/realtek/rtw89/Kconfig | 18 +-
drivers/net/wireless/realtek/rtw89/Makefile | 9 +
drivers/net/wireless/realtek/rtw89/mac.c | 7 +-
drivers/net/wireless/realtek/rtw89/reg.h | 8 +-
drivers/net/wireless/realtek/rtw89/rtw8852c.c | 157 ++++++++++++++++--
.../net/wireless/realtek/rtw89/rtw8852c_rfk.c | 18 ++
.../net/wireless/realtek/rtw89/rtw8852c_rfk.h | 1 +
7 files changed, 203 insertions(+), 15 deletions(-)
--
2.25.1
First one could affect SER because of false alarm event. Second one can
affect spur elimination.
Signed-off-by: Ping-Ke Shih <[email protected]>
---
drivers/net/wireless/realtek/rtw89/reg.h | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/drivers/net/wireless/realtek/rtw89/reg.h b/drivers/net/wireless/realtek/rtw89/reg.h
index dff7992659dce..5c4de043845b2 100644
--- a/drivers/net/wireless/realtek/rtw89/reg.h
+++ b/drivers/net/wireless/realtek/rtw89/reg.h
@@ -2605,7 +2605,6 @@
B_AX_TMAC_HWSIGB_GEN | \
B_AX_TMAC_RXTB | \
B_AX_TMAC_MIMO_CTRL | \
- B_AX_RMAC_CSI | \
B_AX_RMAC_FTM)
#define R_AX_WMAC_TX_TF_INFO_0 0xCCD0
@@ -3667,7 +3666,7 @@
#define R_DCFO 0x4264
#define B_DCFO GENMASK(1, 0)
#define R_SEG0CSI 0x42AC
-#define B_SEG0CSI_IDX GENMASK(10, 0)
+#define B_SEG0CSI_IDX GENMASK(11, 0)
#define R_SEG0CSI_EN 0x42C4
#define B_SEG0CSI_EN BIT(23)
#define R_BSS_CLR_MAP 0x43ac
--
2.25.1
Without this fix, it would cause IOT issue due to AID mismatch.
Signed-off-by: Ping-Ke Shih <[email protected]>
---
drivers/net/wireless/realtek/rtw89/mac.c | 4 ++++
drivers/net/wireless/realtek/rtw89/reg.h | 5 +++++
2 files changed, 9 insertions(+)
diff --git a/drivers/net/wireless/realtek/rtw89/mac.c b/drivers/net/wireless/realtek/rtw89/mac.c
index 07f6634d56a01..a06ca65b339ff 100644
--- a/drivers/net/wireless/realtek/rtw89/mac.c
+++ b/drivers/net/wireless/realtek/rtw89/mac.c
@@ -4240,6 +4240,10 @@ static int rtw89_mac_init_bfee(struct rtw89_dev *rtwdev, u8 mac_idx)
u32_encode_bits(CSI_INIT_RATE_VHT, B_AX_BFMEE_VHT_CSI_RATE_MASK) |
u32_encode_bits(CSI_INIT_RATE_HE, B_AX_BFMEE_HE_CSI_RATE_MASK));
+ reg = rtw89_mac_reg_by_idx(R_AX_CSIRPT_OPTION, mac_idx);
+ rtw89_write32_set(rtwdev, reg,
+ B_AX_CSIPRT_VHTSU_AID_EN | B_AX_CSIPRT_HESU_AID_EN);
+
return 0;
}
diff --git a/drivers/net/wireless/realtek/rtw89/reg.h b/drivers/net/wireless/realtek/rtw89/reg.h
index 6f5d1012c90c6..dff7992659dce 100644
--- a/drivers/net/wireless/realtek/rtw89/reg.h
+++ b/drivers/net/wireless/realtek/rtw89/reg.h
@@ -2842,6 +2842,11 @@
#define R_AX_RX_SR_CTRL_C1 0xEE4A
#define B_AX_SR_EN BIT(0)
+#define R_AX_CSIRPT_OPTION 0xCE64
+#define R_AX_CSIRPT_OPTION_C1 0xEE64
+#define B_AX_CSIPRT_HESU_AID_EN BIT(25)
+#define B_AX_CSIPRT_VHTSU_AID_EN BIT(24)
+
#define R_AX_RX_STATE_MONITOR 0xCEF0
#define R_AX_RX_STATE_MONITOR_C1 0xEEF0
#define B_AX_RX_STATE_MONITOR_MASK GENMASK(31, 0)
--
2.25.1
By experiments, it spends ~45/1090~2480us to enter/leave low power mode,
so the old polling interval 1000us can waste time. Use smaller polling
interval depends on experimental results to reduce the time to transition
state.
Signed-off-by: Ping-Ke Shih <[email protected]>
---
drivers/net/wireless/realtek/rtw89/mac.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/net/wireless/realtek/rtw89/mac.c b/drivers/net/wireless/realtek/rtw89/mac.c
index 05b94842fe662..07f6634d56a01 100644
--- a/drivers/net/wireless/realtek/rtw89/mac.c
+++ b/drivers/net/wireless/realtek/rtw89/mac.c
@@ -1050,6 +1050,7 @@ static int rtw89_mac_check_cpwm_state(struct rtw89_dev *rtwdev,
void rtw89_mac_power_mode_change(struct rtw89_dev *rtwdev, bool enter)
{
enum rtw89_rpwm_req_pwr_state state;
+ unsigned long delay = enter ? 10 : 150;
int ret;
if (enter)
@@ -1059,7 +1060,7 @@ void rtw89_mac_power_mode_change(struct rtw89_dev *rtwdev, bool enter)
rtw89_mac_send_rpwm(rtwdev, state, false);
ret = read_poll_timeout_atomic(rtw89_mac_check_cpwm_state, ret, !ret,
- 1000, 15000, false, rtwdev, state);
+ delay, 15000, false, rtwdev, state);
if (ret)
rtw89_err(rtwdev, "firmware failed to ack for %s ps mode\n",
enter ? "entering" : "leaving");
--
2.25.1