| Message ID | 20230426034737.24870-1-pkshih@realtek.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | c0426c446d92023d344131d01d929bc25db7a24e |
| Delegated to: | Kalle Valo |
| Headers | show |
| Series | [v2] wifi: rtw89: 8852b: adjust quota to avoid SER L1 caused by access null page | expand |
Ping-Ke Shih <pkshih@realtek.com> writes: > Though SER can recover this case, traffic can get stuck for a while. Fix it > by adjusting page quota to avoid hardware access null page of CMAC/DMAC. > > Fixes: a1cb097168fa ("wifi: rtw89: 8852b: configure DLE mem") > Fixes: 3e870b481733 ("wifi: rtw89: 8852b: add HFC quota arrays") > Cc: stable@vger.kernel.org > Tested-by: Larry Finger <Larry.Finger@lwfinger.net> > Link: https://github.com/lwfinger/rtw89/issues/226#issuecomment-1520776761 > Link: https://github.com/lwfinger/rtw89/issues/240 > Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> > --- > v2: add Fixes, Cc and Tested-by tags suggested by Larry. Should this go to wireless tree for v6.4?
> -----Original Message----- > From: Kalle Valo <kvalo@kernel.org> > Sent: Wednesday, April 26, 2023 1:10 PM > To: Ping-Ke Shih <pkshih@realtek.com> > Cc: stable@vger.kernel.org; Larry.Finger@lwfinger.net; linux-wireless@vger.kernel.org > Subject: Re: [PATCH v2] wifi: rtw89: 8852b: adjust quota to avoid SER L1 caused by access null page > > Ping-Ke Shih <pkshih@realtek.com> writes: > > > Though SER can recover this case, traffic can get stuck for a while. Fix it > > by adjusting page quota to avoid hardware access null page of CMAC/DMAC. > > > > Fixes: a1cb097168fa ("wifi: rtw89: 8852b: configure DLE mem") > > Fixes: 3e870b481733 ("wifi: rtw89: 8852b: add HFC quota arrays") > > Cc: stable@vger.kernel.org > > Tested-by: Larry Finger <Larry.Finger@lwfinger.net> > > Link: https://github.com/lwfinger/rtw89/issues/226#issuecomment-1520776761 > > Link: https://github.com/lwfinger/rtw89/issues/240 > > Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> > > --- > > v2: add Fixes, Cc and Tested-by tags suggested by Larry. > > Should this go to wireless tree for v6.4? > Yes, please take it to v6.4. People can get stable connection with this fix. Thank you Ping-Ke
Ping-Ke Shih <pkshih@realtek.com> writes: >> -----Original Message----- >> From: Kalle Valo <kvalo@kernel.org> >> Sent: Wednesday, April 26, 2023 1:10 PM >> To: Ping-Ke Shih <pkshih@realtek.com> >> Cc: stable@vger.kernel.org; Larry.Finger@lwfinger.net; >> linux-wireless@vger.kernel.org >> Subject: Re: [PATCH v2] wifi: rtw89: 8852b: adjust quota to avoid >> SER L1 caused by access null page >> >> Ping-Ke Shih <pkshih@realtek.com> writes: >> >> > Though SER can recover this case, traffic can get stuck for a while. Fix it >> > by adjusting page quota to avoid hardware access null page of CMAC/DMAC. >> > >> > Fixes: a1cb097168fa ("wifi: rtw89: 8852b: configure DLE mem") >> > Fixes: 3e870b481733 ("wifi: rtw89: 8852b: add HFC quota arrays") >> > Cc: stable@vger.kernel.org >> > Tested-by: Larry Finger <Larry.Finger@lwfinger.net> >> > Link: https://github.com/lwfinger/rtw89/issues/226#issuecomment-1520776761 >> > Link: https://github.com/lwfinger/rtw89/issues/240 >> > Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> >> > --- >> > v2: add Fixes, Cc and Tested-by tags suggested by Larry. >> >> Should this go to wireless tree for v6.4? > > Yes, please take it to v6.4. People can get stable connection with this fix. Ok, I'll then take this to wireless.
On 4/26/23 00:16, Ping-Ke Shih wrote: > > >> -----Original Message----- >> From: Kalle Valo <kvalo@kernel.org> >> Sent: Wednesday, April 26, 2023 1:10 PM >> To: Ping-Ke Shih <pkshih@realtek.com> >> Cc: stable@vger.kernel.org; Larry.Finger@lwfinger.net; linux-wireless@vger.kernel.org >> Subject: Re: [PATCH v2] wifi: rtw89: 8852b: adjust quota to avoid SER L1 caused by access null page >> >> Ping-Ke Shih <pkshih@realtek.com> writes: >> >>> Though SER can recover this case, traffic can get stuck for a while. Fix it >>> by adjusting page quota to avoid hardware access null page of CMAC/DMAC. >>> >>> Fixes: a1cb097168fa ("wifi: rtw89: 8852b: configure DLE mem") >>> Fixes: 3e870b481733 ("wifi: rtw89: 8852b: add HFC quota arrays") >>> Cc: stable@vger.kernel.org >>> Tested-by: Larry Finger <Larry.Finger@lwfinger.net> >>> Link: https://github.com/lwfinger/rtw89/issues/226#issuecomment-1520776761 >>> Link: https://github.com/lwfinger/rtw89/issues/240 >>> Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> >>> --- >>> v2: add Fixes, Cc and Tested-by tags suggested by Larry. >> >> Should this go to wireless tree for v6.4? >> > > Yes, please take it to v6.4. People can get stable connection with this fix. I agree. Larry
Ping-Ke Shih <pkshih@realtek.com> wrote: > Though SER can recover this case, traffic can get stuck for a while. Fix it > by adjusting page quota to avoid hardware access null page of CMAC/DMAC. > > Fixes: a1cb097168fa ("wifi: rtw89: 8852b: configure DLE mem") > Fixes: 3e870b481733 ("wifi: rtw89: 8852b: add HFC quota arrays") > Cc: stable@vger.kernel.org > Tested-by: Larry Finger <Larry.Finger@lwfinger.net> > Link: https://github.com/lwfinger/rtw89/issues/226#issuecomment-1520776761 > Link: https://github.com/lwfinger/rtw89/issues/240 > Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> Patch applied to wireless.git, thanks. c0426c446d92 wifi: rtw89: 8852b: adjust quota to avoid SER L1 caused by access null page
diff --git a/drivers/net/wireless/realtek/rtw89/mac.c b/drivers/net/wireless/realtek/rtw89/mac.c index b8019cfc11b20..512de491a064b 100644 --- a/drivers/net/wireless/realtek/rtw89/mac.c +++ b/drivers/net/wireless/realtek/rtw89/mac.c @@ -1425,6 +1425,8 @@ const struct rtw89_mac_size_set rtw89_mac_size = { .wde_size4 = {RTW89_WDE_PG_64, 0, 4096,}, /* PCIE 64 */ .wde_size6 = {RTW89_WDE_PG_64, 512, 0,}, + /* 8852B PCIE SCC */ + .wde_size7 = {RTW89_WDE_PG_64, 510, 2,}, /* DLFW */ .wde_size9 = {RTW89_WDE_PG_64, 0, 1024,}, /* 8852C DLFW */ @@ -1449,6 +1451,8 @@ const struct rtw89_mac_size_set rtw89_mac_size = { .wde_qt4 = {0, 0, 0, 0,}, /* PCIE 64 */ .wde_qt6 = {448, 48, 0, 16,}, + /* 8852B PCIE SCC */ + .wde_qt7 = {446, 48, 0, 16,}, /* 8852C DLFW */ .wde_qt17 = {0, 0, 0, 0,}, /* 8852C PCIE SCC */ diff --git a/drivers/net/wireless/realtek/rtw89/mac.h b/drivers/net/wireless/realtek/rtw89/mac.h index a8d9847ef0b49..6ba633ccdd037 100644 --- a/drivers/net/wireless/realtek/rtw89/mac.h +++ b/drivers/net/wireless/realtek/rtw89/mac.h @@ -792,6 +792,7 @@ struct rtw89_mac_size_set { const struct rtw89_dle_size wde_size0; const struct rtw89_dle_size wde_size4; const struct rtw89_dle_size wde_size6; + const struct rtw89_dle_size wde_size7; const struct rtw89_dle_size wde_size9; const struct rtw89_dle_size wde_size18; const struct rtw89_dle_size wde_size19; @@ -804,6 +805,7 @@ struct rtw89_mac_size_set { const struct rtw89_wde_quota wde_qt0; const struct rtw89_wde_quota wde_qt4; const struct rtw89_wde_quota wde_qt6; + const struct rtw89_wde_quota wde_qt7; const struct rtw89_wde_quota wde_qt17; const struct rtw89_wde_quota wde_qt18; const struct rtw89_ple_quota ple_qt4; diff --git a/drivers/net/wireless/realtek/rtw89/rtw8852b.c b/drivers/net/wireless/realtek/rtw89/rtw8852b.c index eaa2ea0586bc6..6da1b603a9a95 100644 --- a/drivers/net/wireless/realtek/rtw89/rtw8852b.c +++ b/drivers/net/wireless/realtek/rtw89/rtw8852b.c @@ -18,25 +18,25 @@ RTW8852B_FW_BASENAME "-" __stringify(RTW8852B_FW_FORMAT_MAX) ".bin" static const struct rtw89_hfc_ch_cfg rtw8852b_hfc_chcfg_pcie[] = { - {5, 343, grp_0}, /* ACH 0 */ - {5, 343, grp_0}, /* ACH 1 */ - {5, 343, grp_0}, /* ACH 2 */ - {5, 343, grp_0}, /* ACH 3 */ + {5, 341, grp_0}, /* ACH 0 */ + {5, 341, grp_0}, /* ACH 1 */ + {4, 342, grp_0}, /* ACH 2 */ + {4, 342, grp_0}, /* ACH 3 */ {0, 0, grp_0}, /* ACH 4 */ {0, 0, grp_0}, /* ACH 5 */ {0, 0, grp_0}, /* ACH 6 */ {0, 0, grp_0}, /* ACH 7 */ - {4, 344, grp_0}, /* B0MGQ */ - {4, 344, grp_0}, /* B0HIQ */ + {4, 342, grp_0}, /* B0MGQ */ + {4, 342, grp_0}, /* B0HIQ */ {0, 0, grp_0}, /* B1MGQ */ {0, 0, grp_0}, /* B1HIQ */ {40, 0, 0} /* FWCMDQ */ }; static const struct rtw89_hfc_pub_cfg rtw8852b_hfc_pubcfg_pcie = { - 448, /* Group 0 */ + 446, /* Group 0 */ 0, /* Group 1 */ - 448, /* Public Max */ + 446, /* Public Max */ 0 /* WP threshold */ }; @@ -49,13 +49,13 @@ static const struct rtw89_hfc_param_ini rtw8852b_hfc_param_ini_pcie[] = { }; static const struct rtw89_dle_mem rtw8852b_dle_mem_pcie[] = { - [RTW89_QTA_SCC] = {RTW89_QTA_SCC, &rtw89_mac_size.wde_size6, - &rtw89_mac_size.ple_size6, &rtw89_mac_size.wde_qt6, - &rtw89_mac_size.wde_qt6, &rtw89_mac_size.ple_qt18, + [RTW89_QTA_SCC] = {RTW89_QTA_SCC, &rtw89_mac_size.wde_size7, + &rtw89_mac_size.ple_size6, &rtw89_mac_size.wde_qt7, + &rtw89_mac_size.wde_qt7, &rtw89_mac_size.ple_qt18, &rtw89_mac_size.ple_qt58}, - [RTW89_QTA_WOW] = {RTW89_QTA_WOW, &rtw89_mac_size.wde_size6, - &rtw89_mac_size.ple_size6, &rtw89_mac_size.wde_qt6, - &rtw89_mac_size.wde_qt6, &rtw89_mac_size.ple_qt18, + [RTW89_QTA_WOW] = {RTW89_QTA_WOW, &rtw89_mac_size.wde_size7, + &rtw89_mac_size.ple_size6, &rtw89_mac_size.wde_qt7, + &rtw89_mac_size.wde_qt7, &rtw89_mac_size.ple_qt18, &rtw89_mac_size.ple_qt_52b_wow}, [RTW89_QTA_DLFW] = {RTW89_QTA_DLFW, &rtw89_mac_size.wde_size9, &rtw89_mac_size.ple_size8, &rtw89_mac_size.wde_qt4,
Though SER can recover this case, traffic can get stuck for a while. Fix it by adjusting page quota to avoid hardware access null page of CMAC/DMAC. Fixes: a1cb097168fa ("wifi: rtw89: 8852b: configure DLE mem") Fixes: 3e870b481733 ("wifi: rtw89: 8852b: add HFC quota arrays") Cc: stable@vger.kernel.org Tested-by: Larry Finger <Larry.Finger@lwfinger.net> Link: https://github.com/lwfinger/rtw89/issues/226#issuecomment-1520776761 Link: https://github.com/lwfinger/rtw89/issues/240 Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> --- v2: add Fixes, Cc and Tested-by tags suggested by Larry. --- drivers/net/wireless/realtek/rtw89/mac.c | 4 +++ drivers/net/wireless/realtek/rtw89/mac.h | 2 ++ drivers/net/wireless/realtek/rtw89/rtw8852b.c | 28 +++++++++---------- 3 files changed, 20 insertions(+), 14 deletions(-)