| Message ID | fa9ff2b4-6e1a-45e4-90de-db6fe0d4f433@gmail.com (mailing list archive) |
|---|---|
| State | New |
| Delegated to: | Ping-Ke Shih |
| Headers | show |
| Series | wifi: rtw88: Prepare to support RTL8814AU (part 1/2) | expand |
Bitterblue Smith <rtl8821cerfe2@gmail.com> wrote: > The existing code is suitable for chips with up to 2 spatial streams. > Inform the firmware about the rates it's allowed to use when > transmitting 3 spatial streams. > > Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com> > --- > drivers/net/wireless/realtek/rtw88/fw.c | 14 ++++++++++++++ > drivers/net/wireless/realtek/rtw88/fw.h | 1 + > 2 files changed, 15 insertions(+) > > diff --git a/drivers/net/wireless/realtek/rtw88/fw.c b/drivers/net/wireless/realtek/rtw88/fw.c > index 02389b7c6876..0ca1b139110d 100644 > --- a/drivers/net/wireless/realtek/rtw88/fw.c > +++ b/drivers/net/wireless/realtek/rtw88/fw.c > @@ -735,6 +735,7 @@ void rtw_fw_send_ra_info(struct rtw_dev *rtwdev, struct rtw_sta_info *si, > { > u8 h2c_pkt[H2C_PKT_SIZE] = {0}; > bool disable_pt = true; > + u32 mask_hi; > > SET_H2C_CMD_ID_CLASS(h2c_pkt, H2C_CMD_RA_INFO); > > @@ -755,6 +756,19 @@ void rtw_fw_send_ra_info(struct rtw_dev *rtwdev, struct rtw_sta_info *si, > si->init_ra_lv = 0; > > rtw_fw_send_h2c_command(rtwdev, h2c_pkt); > + > + if (rtwdev->chip->rf_tbl[RF_PATH_C]) { Using `efuse->hw_cap.nss >= 3` would be consistent with latter patch. > + SET_H2C_CMD_ID_CLASS(h2c_pkt, H2C_CMD_RA_INFO_HI); > + > + mask_hi = si->ra_mask >> 32; > + > + SET_RA_INFO_RA_MASK0(h2c_pkt, (mask_hi & 0xff)); > + SET_RA_INFO_RA_MASK1(h2c_pkt, (mask_hi & 0xff00) >> 8); > + SET_RA_INFO_RA_MASK2(h2c_pkt, (mask_hi & 0xff0000) >> 16); > + SET_RA_INFO_RA_MASK3(h2c_pkt, (mask_hi & 0xff000000) >> 24); > + > + rtw_fw_send_h2c_command(rtwdev, h2c_pkt); > + } > } Prefer calling RA_INFO LO/HI in the same level. I meant rtw_fw_send_ra_info() { rtw_fw_send_ra_info_lo(); // original RA info if (efuse->hw_cap.nss <= 2) return; rtw_fw_send_ra_info_hi(); }
On 27/01/2025 08:36, Ping-Ke Shih wrote: > Bitterblue Smith <rtl8821cerfe2@gmail.com> wrote: >> The existing code is suitable for chips with up to 2 spatial streams. >> Inform the firmware about the rates it's allowed to use when >> transmitting 3 spatial streams. >> >> Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com> >> --- >> drivers/net/wireless/realtek/rtw88/fw.c | 14 ++++++++++++++ >> drivers/net/wireless/realtek/rtw88/fw.h | 1 + >> 2 files changed, 15 insertions(+) >> >> diff --git a/drivers/net/wireless/realtek/rtw88/fw.c b/drivers/net/wireless/realtek/rtw88/fw.c >> index 02389b7c6876..0ca1b139110d 100644 >> --- a/drivers/net/wireless/realtek/rtw88/fw.c >> +++ b/drivers/net/wireless/realtek/rtw88/fw.c >> @@ -735,6 +735,7 @@ void rtw_fw_send_ra_info(struct rtw_dev *rtwdev, struct rtw_sta_info *si, >> { >> u8 h2c_pkt[H2C_PKT_SIZE] = {0}; >> bool disable_pt = true; >> + u32 mask_hi; >> >> SET_H2C_CMD_ID_CLASS(h2c_pkt, H2C_CMD_RA_INFO); >> >> @@ -755,6 +756,19 @@ void rtw_fw_send_ra_info(struct rtw_dev *rtwdev, struct rtw_sta_info *si, >> si->init_ra_lv = 0; >> >> rtw_fw_send_h2c_command(rtwdev, h2c_pkt); >> + >> + if (rtwdev->chip->rf_tbl[RF_PATH_C]) { > > Using `efuse->hw_cap.nss >= 3` would be consistent with latter patch. > I would like that, but nss is 2 when RTL8814AU is in USB 2 mode. I assume this is to keep the current draw under the 500 mA limit of USB 2. What about rtwdev->hal.rf_path_num >= 3 ? I don't remember why I didn't do that. >> + SET_H2C_CMD_ID_CLASS(h2c_pkt, H2C_CMD_RA_INFO_HI); >> + >> + mask_hi = si->ra_mask >> 32; >> + >> + SET_RA_INFO_RA_MASK0(h2c_pkt, (mask_hi & 0xff)); >> + SET_RA_INFO_RA_MASK1(h2c_pkt, (mask_hi & 0xff00) >> 8); >> + SET_RA_INFO_RA_MASK2(h2c_pkt, (mask_hi & 0xff0000) >> 16); >> + SET_RA_INFO_RA_MASK3(h2c_pkt, (mask_hi & 0xff000000) >> 24); >> + >> + rtw_fw_send_h2c_command(rtwdev, h2c_pkt); >> + } >> } > > Prefer calling RA_INFO LO/HI in the same level. I meant > > rtw_fw_send_ra_info() > { > rtw_fw_send_ra_info_lo(); // original RA info > > if (efuse->hw_cap.nss <= 2) > return; > > rtw_fw_send_ra_info_hi(); > } >
Bitterblue Smith <rtl8821cerfe2@gmail.com> wrote: > On 27/01/2025 08:36, Ping-Ke Shih wrote: > > Bitterblue Smith <rtl8821cerfe2@gmail.com> wrote: > >> The existing code is suitable for chips with up to 2 spatial streams. > >> Inform the firmware about the rates it's allowed to use when > >> transmitting 3 spatial streams. > >> > >> Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com> > >> --- > >> drivers/net/wireless/realtek/rtw88/fw.c | 14 ++++++++++++++ > >> drivers/net/wireless/realtek/rtw88/fw.h | 1 + > >> 2 files changed, 15 insertions(+) > >> > >> diff --git a/drivers/net/wireless/realtek/rtw88/fw.c b/drivers/net/wireless/realtek/rtw88/fw.c > >> index 02389b7c6876..0ca1b139110d 100644 > >> --- a/drivers/net/wireless/realtek/rtw88/fw.c > >> +++ b/drivers/net/wireless/realtek/rtw88/fw.c > >> @@ -735,6 +735,7 @@ void rtw_fw_send_ra_info(struct rtw_dev *rtwdev, struct rtw_sta_info *si, > >> { > >> u8 h2c_pkt[H2C_PKT_SIZE] = {0}; > >> bool disable_pt = true; > >> + u32 mask_hi; > >> > >> SET_H2C_CMD_ID_CLASS(h2c_pkt, H2C_CMD_RA_INFO); > >> > >> @@ -755,6 +756,19 @@ void rtw_fw_send_ra_info(struct rtw_dev *rtwdev, struct rtw_sta_info *si, > >> si->init_ra_lv = 0; > >> > >> rtw_fw_send_h2c_command(rtwdev, h2c_pkt); > >> + > >> + if (rtwdev->chip->rf_tbl[RF_PATH_C]) { > > > > Using `efuse->hw_cap.nss >= 3` would be consistent with latter patch. > > > > I would like that, but nss is 2 when RTL8814AU is in USB 2 mode. > I assume this is to keep the current draw under the 500 mA limit > of USB 2. > > What about rtwdev->hal.rf_path_num >= 3 ? I don't remember why > I didn't do that. I think `rtwdev->hal.rf_path_num >= 3` is suitable to initialize/configure hardware registers, because no matter USB 2 or 3 mode should be the same. For this case (RA info), this is related to protocol, so I feel `efuse->hw_cap.nss >= 3` is suitable, but I have not seen a patch to declare supported NSS in register_hw(), or I missed it? Or, without RA_INFO_HI, it gets abnormal rate to RTL8814AU in your test?
On 28/01/2025 07:52, Ping-Ke Shih wrote: > Bitterblue Smith <rtl8821cerfe2@gmail.com> wrote: >> On 27/01/2025 08:36, Ping-Ke Shih wrote: >>> Bitterblue Smith <rtl8821cerfe2@gmail.com> wrote: >>>> The existing code is suitable for chips with up to 2 spatial streams. >>>> Inform the firmware about the rates it's allowed to use when >>>> transmitting 3 spatial streams. >>>> >>>> Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com> >>>> --- >>>> drivers/net/wireless/realtek/rtw88/fw.c | 14 ++++++++++++++ >>>> drivers/net/wireless/realtek/rtw88/fw.h | 1 + >>>> 2 files changed, 15 insertions(+) >>>> >>>> diff --git a/drivers/net/wireless/realtek/rtw88/fw.c b/drivers/net/wireless/realtek/rtw88/fw.c >>>> index 02389b7c6876..0ca1b139110d 100644 >>>> --- a/drivers/net/wireless/realtek/rtw88/fw.c >>>> +++ b/drivers/net/wireless/realtek/rtw88/fw.c >>>> @@ -735,6 +735,7 @@ void rtw_fw_send_ra_info(struct rtw_dev *rtwdev, struct rtw_sta_info *si, >>>> { >>>> u8 h2c_pkt[H2C_PKT_SIZE] = {0}; >>>> bool disable_pt = true; >>>> + u32 mask_hi; >>>> >>>> SET_H2C_CMD_ID_CLASS(h2c_pkt, H2C_CMD_RA_INFO); >>>> >>>> @@ -755,6 +756,19 @@ void rtw_fw_send_ra_info(struct rtw_dev *rtwdev, struct rtw_sta_info *si, >>>> si->init_ra_lv = 0; >>>> >>>> rtw_fw_send_h2c_command(rtwdev, h2c_pkt); >>>> + >>>> + if (rtwdev->chip->rf_tbl[RF_PATH_C]) { >>> >>> Using `efuse->hw_cap.nss >= 3` would be consistent with latter patch. >>> >> >> I would like that, but nss is 2 when RTL8814AU is in USB 2 mode. >> I assume this is to keep the current draw under the 500 mA limit >> of USB 2. >> >> What about rtwdev->hal.rf_path_num >= 3 ? I don't remember why >> I didn't do that. > > I think `rtwdev->hal.rf_path_num >= 3` is suitable to initialize/configure > hardware registers, because no matter USB 2 or 3 mode should be the same. > > For this case (RA info), this is related to protocol, so I feel > `efuse->hw_cap.nss >= 3` is suitable, but I have not seen a patch to declare > supported NSS in register_hw(), or I missed it? Or, without RA_INFO_HI, > it gets abnormal rate to RTL8814AU in your test? > You didn't miss it, that will be in part 3. You can see the code here: https://github.com/lwfinger/rtw88/blob/21a3fa7ec11a0cbb3be14145f45cdca35c3d3217/rtw8814a.c#L82 I get a steady 300 Mbps even without RA_INFO_HI. The problem is only cosmetic. Right after connecting to the AP the first C2H_RA_RPT comes a bit later than usual, after the first CTRL-EVENT-SIGNAL-CHANGE, so we see a TX rate of 0: Jan 28 19:00:06 ideapad2 NetworkManager[504]: <info> [1738083606.4996] manager: NetworkManager state is now CONNECTED_SITE Jan 28 19:00:06 ideapad2 NetworkManager[504]: <info> [1738083606.5003] device (wlp3s0f3u2): Activation: successful, device activated. Jan 28 19:00:06 ideapad2 wpa_supplicant[1480]: wlp3s0f3u2: CTRL-EVENT-SIGNAL-CHANGE above=1 signal=-42 noise=9999 txrate=0 Jan 28 19:00:06 ideapad2 NetworkManager[504]: <info> [1738083606.6020] manager: NetworkManager state is now CONNECTED_GLOBAL With RA_INFO_HI the first C2H_RA_RPT comes at the normal time, before the first CTRL-EVENT-SIGNAL-CHANGE: Jan 28 19:09:44 ideapad2 NetworkManager[504]: <info> [1738084184.2400] manager: NetworkManager state is now CONNECTED_SITE Jan 28 19:09:44 ideapad2 NetworkManager[504]: <info> [1738084184.2407] device (wlp3s0f3u2): Activation: successful, device activated. Jan 28 19:09:44 ideapad2 wpa_supplicant[1480]: wlp3s0f3u2: CTRL-EVENT-SIGNAL-CHANGE above=1 signal=-44 noise=9999 txrate=780000 Jan 28 19:09:44 ideapad2 NetworkManager[504]: <info> [1738084184.3810] manager: NetworkManager state is now CONNECTED_GLOBAL
Bitterblue Smith <rtl8821cerfe2@gmail.com> wrote: > On 28/01/2025 07:52, Ping-Ke Shih wrote: > > Bitterblue Smith <rtl8821cerfe2@gmail.com> wrote: > >> On 27/01/2025 08:36, Ping-Ke Shih wrote: > >>> Bitterblue Smith <rtl8821cerfe2@gmail.com> wrote: > >>>> The existing code is suitable for chips with up to 2 spatial streams. > >>>> Inform the firmware about the rates it's allowed to use when > >>>> transmitting 3 spatial streams. > >>>> > >>>> Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com> > >>>> --- > >>>> drivers/net/wireless/realtek/rtw88/fw.c | 14 ++++++++++++++ > >>>> drivers/net/wireless/realtek/rtw88/fw.h | 1 + > >>>> 2 files changed, 15 insertions(+) > >>>> > >>>> diff --git a/drivers/net/wireless/realtek/rtw88/fw.c b/drivers/net/wireless/realtek/rtw88/fw.c > >>>> index 02389b7c6876..0ca1b139110d 100644 > >>>> --- a/drivers/net/wireless/realtek/rtw88/fw.c > >>>> +++ b/drivers/net/wireless/realtek/rtw88/fw.c > >>>> @@ -735,6 +735,7 @@ void rtw_fw_send_ra_info(struct rtw_dev *rtwdev, struct rtw_sta_info *si, > >>>> { > >>>> u8 h2c_pkt[H2C_PKT_SIZE] = {0}; > >>>> bool disable_pt = true; > >>>> + u32 mask_hi; > >>>> > >>>> SET_H2C_CMD_ID_CLASS(h2c_pkt, H2C_CMD_RA_INFO); > >>>> > >>>> @@ -755,6 +756,19 @@ void rtw_fw_send_ra_info(struct rtw_dev *rtwdev, struct rtw_sta_info *si, > >>>> si->init_ra_lv = 0; > >>>> > >>>> rtw_fw_send_h2c_command(rtwdev, h2c_pkt); > >>>> + > >>>> + if (rtwdev->chip->rf_tbl[RF_PATH_C]) { > >>> > >>> Using `efuse->hw_cap.nss >= 3` would be consistent with latter patch. > >>> > >> > >> I would like that, but nss is 2 when RTL8814AU is in USB 2 mode. > >> I assume this is to keep the current draw under the 500 mA limit > >> of USB 2. > >> > >> What about rtwdev->hal.rf_path_num >= 3 ? I don't remember why > >> I didn't do that. > > > > I think `rtwdev->hal.rf_path_num >= 3` is suitable to initialize/configure > > hardware registers, because no matter USB 2 or 3 mode should be the same. > > > > For this case (RA info), this is related to protocol, so I feel > > `efuse->hw_cap.nss >= 3` is suitable, but I have not seen a patch to declare > > supported NSS in register_hw(), or I missed it? Or, without RA_INFO_HI, > > it gets abnormal rate to RTL8814AU in your test? > > > > You didn't miss it, that will be in part 3. You can see the code here: > > https://github.com/lwfinger/rtw88/blob/21a3fa7ec11a0cbb3be14145f45cdca35c3d3217/rtw8814a.c#L82 > I feel we should clearly define the meaning. What I thought for 8814AU are: - hal->rf_type: hardware capability. Should be RF_3T3R no matter USB 2 or 3. - hal->antenna_tx: the antenna for current TX. Can be 2 antenna. - hal->antenna_rx: the antenna for current RX. Can be 2 antenna. - efuse->hw_cap.nss: read from efuse. So this will be 3SS for USB 2/3. If you have better defnitiion, please share your ideas. Also, I would want to point the mcs_map implemented in rtw_init_vht_cap(). if (efuse->hw_cap.nss > 1) { highest = cpu_to_le16(780); mcs_map |= IEEE80211_VHT_MCS_SUPPORT_0_9 << 2; } This can only support up to 2SS, and I'm not sure if 'efuse->hw_cap.nss' is still sutiable for 8814AU. > > With RA_INFO_HI the first C2H_RA_RPT comes at the normal time, > before the first CTRL-EVENT-SIGNAL-CHANGE: So, how about sending RA_INFO_HI unconditionally for 8814AU (just check chip ID as condition)?
diff --git a/drivers/net/wireless/realtek/rtw88/fw.c b/drivers/net/wireless/realtek/rtw88/fw.c index 02389b7c6876..0ca1b139110d 100644 --- a/drivers/net/wireless/realtek/rtw88/fw.c +++ b/drivers/net/wireless/realtek/rtw88/fw.c @@ -735,6 +735,7 @@ void rtw_fw_send_ra_info(struct rtw_dev *rtwdev, struct rtw_sta_info *si, { u8 h2c_pkt[H2C_PKT_SIZE] = {0}; bool disable_pt = true; + u32 mask_hi; SET_H2C_CMD_ID_CLASS(h2c_pkt, H2C_CMD_RA_INFO); @@ -755,6 +756,19 @@ void rtw_fw_send_ra_info(struct rtw_dev *rtwdev, struct rtw_sta_info *si, si->init_ra_lv = 0; rtw_fw_send_h2c_command(rtwdev, h2c_pkt); + + if (rtwdev->chip->rf_tbl[RF_PATH_C]) { + SET_H2C_CMD_ID_CLASS(h2c_pkt, H2C_CMD_RA_INFO_HI); + + mask_hi = si->ra_mask >> 32; + + SET_RA_INFO_RA_MASK0(h2c_pkt, (mask_hi & 0xff)); + SET_RA_INFO_RA_MASK1(h2c_pkt, (mask_hi & 0xff00) >> 8); + SET_RA_INFO_RA_MASK2(h2c_pkt, (mask_hi & 0xff0000) >> 16); + SET_RA_INFO_RA_MASK3(h2c_pkt, (mask_hi & 0xff000000) >> 24); + + rtw_fw_send_h2c_command(rtwdev, h2c_pkt); + } } void rtw_fw_media_status_report(struct rtw_dev *rtwdev, u8 mac_id, bool connect) diff --git a/drivers/net/wireless/realtek/rtw88/fw.h b/drivers/net/wireless/realtek/rtw88/fw.h index 404de1b0c407..48ad9ceab6ea 100644 --- a/drivers/net/wireless/realtek/rtw88/fw.h +++ b/drivers/net/wireless/realtek/rtw88/fw.h @@ -557,6 +557,7 @@ static inline void rtw_h2c_pkt_set_header(u8 *h2c_pkt, u8 sub_id) #define H2C_CMD_DEFAULT_PORT 0x2c #define H2C_CMD_RA_INFO 0x40 #define H2C_CMD_RSSI_MONITOR 0x42 +#define H2C_CMD_RA_INFO_HI 0x46 #define H2C_CMD_BCN_FILTER_OFFLOAD_P0 0x56 #define H2C_CMD_BCN_FILTER_OFFLOAD_P1 0x57 #define H2C_CMD_WL_PHY_INFO 0x58
The existing code is suitable for chips with up to 2 spatial streams. Inform the firmware about the rates it's allowed to use when transmitting 3 spatial streams. Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com> --- drivers/net/wireless/realtek/rtw88/fw.c | 14 ++++++++++++++ drivers/net/wireless/realtek/rtw88/fw.h | 1 + 2 files changed, 15 insertions(+)