From: Mehmet Fide <mehmet.fide@gmail.com>
To: Ping-Ke Shih <pkshih@realtek.com>
Cc: Luka Gejak <luka.gejak@linux.dev>,
Bitterblue Smith <rtl8821cerfe2@gmail.com>,
linux-wireless@vger.kernel.org, linux-kernel@vger.kernel.org,
mehmet.fide@screeningeagle.com
Subject: [PATCH rtw-next v10 3/3] wifi: rtw88: time the countdown downloads from the TBTT
Date: Wed, 11 Nov 2026 13:15:25 +0100 [thread overview]
Message-ID: <20261111121525.358041-4-mehmet.fide@gmail.com> (raw)
In-Reply-To: <20261111121525.358041-1-mehmet.fide@gmail.com>
From: Mehmet Fide <mehmet.fide@screeningeagle.com>
The countdown downloads are queued one beacon interval apart, with no
relation to the TBTT. A download that runs late past the next TBTT
leaves a count out of the air.
Queue each download 10 ms after the next TBTT, read from the TSF, so it
may run late by almost a beacon interval; a later one is followed by
the next at once. A TSF jump, an IBSS merge or a reset, restarts the
schedule from the next TBTT.
The first count is downloaded the same way, not when mac80211 changes
the beacon: a download close to a TBTT keeps beacons off the air. On an
RTL8821CU and an RTL8822CU a TBTT during the download sent no beacon or
the old one. On an RTL8822CE a TBTT up to about 100 us before the
download sent none and the next beacon stayed out too, and with a count
of 2 no CSA beacon went out. The switch completes up to one beacon
interval later. The replay after a hardware restart still downloads at
once.
The USB chips raise no TBTT interrupt, so the TSF register is read. The
beaconing interface is always on port 0: rtw_core_port_switch() moves
an AP there, an IBSS runs alone.
Signed-off-by: Mehmet Fide <mehmet.fide@screeningeagle.com>
---
Notes:
Changes in v10:
- the first count is downloaded by the countdown work, 10 ms after the
next TBTT, and no longer when mac80211 changes the beacon. v9 took a
TBTT up to 10 ms before the start as the one that sent the first
count, but a TBTT during the download sends no CSA beacon, and with
a count of 2 none went out at all (Luka). No download runs close to
a TBTT now, so there is nothing left to guess
- rebased on rtw-next
- Acked-by from Ping-Ke and Reviewed-by from Luka dropped for the
changed code
Changes in v9:
- Acked-by from Ping-Ke (on v8) and Reviewed-by from Luka
Changes in v8:
- the TSF read retries once, if the high word changed, instead of
looping (Ping-Ke)
- rtw_fw_csa_queue() takes a start argument and reads the TSF once,
also on start (Ping-Ke)
- the port 0 comment is gone, the commit message says it (Ping-Ke)
- empty lines where asked (Ping-Ke)
Changes in v7:
- new patch, from Ping-Ke's review of the countdown pacing in v6
drivers/net/wireless/realtek/rtw88/fw.c | 48 ++++++++++++++++---
drivers/net/wireless/realtek/rtw88/mac80211.c | 7 ++-
drivers/net/wireless/realtek/rtw88/main.h | 2 +-
drivers/net/wireless/realtek/rtw88/reg.h | 1 +
4 files changed, 48 insertions(+), 10 deletions(-)
diff --git a/drivers/net/wireless/realtek/rtw88/fw.c b/drivers/net/wireless/realtek/rtw88/fw.c
index 2d81aaeed0d7..c30aa938f9a9 100644
--- a/drivers/net/wireless/realtek/rtw88/fw.c
+++ b/drivers/net/wireless/realtek/rtw88/fw.c
@@ -3,6 +3,7 @@
*/
#include <linux/iopoll.h>
+#include <linux/math64.h>
#include "main.h"
#include "coex.h"
@@ -1809,14 +1810,47 @@ bool rtw_fw_csa_active(struct rtw_dev *rtwdev)
return !!rtwdev->csa_vif;
}
-/* paced from the first download: a late one does not delay the next */
-static void rtw_fw_csa_queue(struct rtw_dev *rtwdev, struct ieee80211_vif *vif)
+#define RTW_CSA_TBTT_MARGIN_US 10000
+
+static u64 rtw_fw_read_tsf(struct rtw_dev *rtwdev)
+{
+ u32 hi, lo, tmp;
+
+ hi = rtw_read32(rtwdev, REG_TSFTR + 4);
+ lo = rtw_read32(rtwdev, REG_TSFTR);
+ tmp = rtw_read32(rtwdev, REG_TSFTR + 4);
+ if (tmp != hi) {
+ hi = tmp;
+ lo = rtw_read32(rtwdev, REG_TSFTR);
+ }
+
+ return ((u64)hi << 32) | lo;
+}
+
+static void rtw_fw_csa_queue(struct rtw_dev *rtwdev, struct ieee80211_vif *vif,
+ bool start)
{
u32 interval = ieee80211_tu_to_usec(vif->bss_conf.beacon_int);
+ u64 tsf = rtw_fw_read_tsf(rtwdev);
s64 delay;
+ u32 rem;
+
+ if (start) {
+ div_u64_rem(tsf, interval, &rem);
+ rtwdev->csa_tbtt = tsf - rem + interval;
+ } else {
+ rtwdev->csa_tbtt += interval;
+ }
+
+ delay = rtwdev->csa_tbtt + RTW_CSA_TBTT_MARGIN_US - tsf;
+ if (delay > interval + RTW_CSA_TBTT_MARGIN_US ||
+ delay < -(s64)interval) {
+ /* the TSF jumped: restart from the next TBTT */
+ div_u64_rem(tsf, interval, &rem);
+ rtwdev->csa_tbtt = tsf - rem + interval;
+ delay = rtwdev->csa_tbtt + RTW_CSA_TBTT_MARGIN_US - tsf;
+ }
- rtwdev->csa_due = ktime_add_us(rtwdev->csa_due, interval);
- delay = ktime_us_delta(rtwdev->csa_due, ktime_get());
ieee80211_queue_delayed_work(rtwdev->hw, &rtwdev->csa_beacon_work,
usecs_to_jiffies(max(delay, 0)));
}
@@ -1829,8 +1863,8 @@ void rtw_fw_csa_start(struct rtw_dev *rtwdev, struct ieee80211_vif *vif)
return;
rtwdev->csa_vif = vif;
- rtwdev->csa_due = ktime_get();
- rtw_fw_csa_queue(rtwdev, vif);
+
+ rtw_fw_csa_queue(rtwdev, vif, true);
}
void rtw_fw_csa_stop(struct rtw_dev *rtwdev, struct ieee80211_vif *vif)
@@ -1880,7 +1914,7 @@ void rtw_fw_csa_beacon_work(struct work_struct *work)
if (!ieee80211_beacon_cntdwn_is_complete(vif, 0)) {
rtw_fw_download_rsvd_page(rtwdev);
rtw_send_rsvd_page_h2c(rtwdev);
- rtw_fw_csa_queue(rtwdev, vif);
+ rtw_fw_csa_queue(rtwdev, vif, false);
} else {
ieee80211_csa_finish(vif, 0);
}
diff --git a/drivers/net/wireless/realtek/rtw88/mac80211.c b/drivers/net/wireless/realtek/rtw88/mac80211.c
index e6c9866fd599..dbd148760b9b 100644
--- a/drivers/net/wireless/realtek/rtw88/mac80211.c
+++ b/drivers/net/wireless/realtek/rtw88/mac80211.c
@@ -441,8 +441,11 @@ static void rtw_ops_bss_info_changed(struct ieee80211_hw *hw,
if (changed & BSS_CHANGED_BEACON) {
rtw_set_dtim_period(rtwdev, conf->dtim_period);
- rtw_fw_download_rsvd_page(rtwdev);
- rtw_send_rsvd_page_h2c(rtwdev);
+ /* the countdown downloads its beacons after a TBTT */
+ if (!conf->csa_active || changed & BSS_CHANGED_BEACON_ENABLED) {
+ rtw_fw_download_rsvd_page(rtwdev);
+ rtw_send_rsvd_page_h2c(rtwdev);
+ }
/* restart replay resumes it, a beacon without CSA ends it */
if (conf->csa_active)
rtw_fw_csa_start(rtwdev, vif);
diff --git a/drivers/net/wireless/realtek/rtw88/main.h b/drivers/net/wireless/realtek/rtw88/main.h
index 5d18841a705f..466992b85c43 100644
--- a/drivers/net/wireless/realtek/rtw88/main.h
+++ b/drivers/net/wireless/realtek/rtw88/main.h
@@ -2096,7 +2096,7 @@ struct rtw_dev {
struct work_struct update_beacon_work;
struct delayed_work csa_beacon_work;
struct ieee80211_vif *csa_vif;
- ktime_t csa_due;
+ u64 csa_tbtt;
/* used to protect txqs list */
spinlock_t txq_lock;
diff --git a/drivers/net/wireless/realtek/rtw88/reg.h b/drivers/net/wireless/realtek/rtw88/reg.h
index afcdb66fdb9a..3ff4ae0885ef 100644
--- a/drivers/net/wireless/realtek/rtw88/reg.h
+++ b/drivers/net/wireless/realtek/rtw88/reg.h
@@ -506,6 +506,7 @@
#define REG_USTIME_TSF 0x055C
#define REG_BCN_MAX_ERR 0x055D
#define REG_RXTSF_OFFSET_CCK 0x055E
+#define REG_TSFTR 0x0560
#define REG_MISC_CTRL 0x0577
#define BIT_EN_FREE_CNT BIT(3)
#define BIT_DIS_SECOND_CCA (BIT(0) | BIT(1))
--
2.56.0
prev parent reply other threads:[~2026-11-11 12:15 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-11 12:15 [PATCH rtw-next v10 0/3] wifi: rtw88: channel switch in AP mode Mehmet Fide
2026-11-11 12:15 ` [PATCH rtw-next v10 1/3] wifi: rtw88: download the beacon the reserved page was built with Mehmet Fide
2026-11-11 12:15 ` [PATCH rtw-next v10 2/3] wifi: rtw88: support channel switch in AP mode Mehmet Fide
2026-11-11 12:15 ` Mehmet Fide [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261111121525.358041-4-mehmet.fide@gmail.com \
--to=mehmet.fide@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-wireless@vger.kernel.org \
--cc=luka.gejak@linux.dev \
--cc=mehmet.fide@screeningeagle.com \
--cc=pkshih@realtek.com \
--cc=rtl8821cerfe2@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®