From: Felix Fietkau <nbd@nbd.name>
To: linux-wireless@vger.kernel.org
Subject: [PATCH mt76-next 10/20] wifi: mt76: usb: do not park the TX worker again under MT76_RESET
Date: Fri, 20 Nov 2026 18:44:49 +0100 [thread overview]
Message-ID: <20261120174459.1657382-10-nbd@nbd.name> (raw)
In-Reply-To: <20261120174459.1657382-1-nbd@nbd.name>
When the wait for pending TX in mt76u_stop_tx() times out, its cleanup
parks the TX worker also where the caller parked it already. This is the
case for mt792xu_disconnect(), which reaches mt76u_stop_tx() through
mt76_unregister_device(), and for mt7921u_mac_reset() and
mt7925u_mac_reset(). The second park triggers a WARN, and the cleanup
then unparks the worker, which runs again during the teardown or the
chip reset:
mt7925u 2-4:1.0: timed out waiting for pending tx
WARNING: CPU: 1 PID: 15805 at kthread_park+0x73/0x90
Call Trace:
...
mt792xu_stop+0x15/0x30 [mt792x_usb]
...
drv_stop+0x33/0xd0 [mac80211]
...
ieee80211_unregister_hw+0x45/0x140 [mac80211]
mt76_unregister_device+0x70/0xf0 [mt76]
mt792xu_disconnect+0x7a/0x1c0 [mt792x_usb]
...
Park and unpark the worker in the cleanup only while MT76_RESET is
clear, as mt792x_flush() does. Every caller that holds MT76_RESET has
parked the worker already, or runs before mt76_register_device() starts
it.
Fixes: 7910bd565d8b ("wifi: mt76: mt792x: quiesce USB paths on disconnect")
Fixes: 0d2afe09fad5 ("mt76: mt7921: add mt7921u driver")
Fixes: c948b5da6bbe ("wifi: mt76: mt7925: add Mediatek Wi-Fi7 driver for mt7925 chips")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
---
drivers/net/wireless/mediatek/mt76/usb.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/net/wireless/mediatek/mt76/usb.c b/drivers/net/wireless/mediatek/mt76/usb.c
index be48a88a3ad0..9a701840d23a 100644
--- a/drivers/net/wireless/mediatek/mt76/usb.c
+++ b/drivers/net/wireless/mediatek/mt76/usb.c
@@ -1070,6 +1070,7 @@ void mt76u_stop_tx(struct mt76_dev *dev)
if (!ret) {
struct mt76_queue_entry entry;
struct mt76_queue *q;
+ bool park;
int i, j;
dev_err(dev->dev, "timed out waiting for pending tx\n");
@@ -1083,7 +1084,9 @@ void mt76u_stop_tx(struct mt76_dev *dev)
usb_kill_urb(q->entry[j].urb);
}
- mt76_worker_disable(&dev->tx_worker);
+ park = !test_bit(MT76_RESET, &dev->phy.state);
+ if (park)
+ mt76_worker_disable(&dev->tx_worker);
/* On device removal we maight queue skb's, but mt76u_tx_kick()
* will fail to submit urb, cleanup those skb's manually.
@@ -1100,7 +1103,8 @@ void mt76u_stop_tx(struct mt76_dev *dev)
}
}
- mt76_worker_enable(&dev->tx_worker);
+ if (park)
+ mt76_worker_enable(&dev->tx_worker);
}
cancel_work_sync(&dev->usb.stat_work);
--
2.47.3
next prev parent reply other threads:[~2026-11-20 17:45 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-20 17:44 [PATCH mt76-next 01/20] wifi: mt76: mt7925: free the vif links when all links go Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 02/20] wifi: mt76: mt7925: free removed vif links after an RCU grace period Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 03/20] wifi: mt76: remove a replayed station from the driver before the add Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 04/20] wifi: mt76: mt7925: run the reset link replay under the wiphy lock Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 05/20] wifi: mt76: mt7925: take the mutex before the link_info_changed lookup Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 06/20] wifi: mt76: mt7925: hold the mutex for the link lookup in conf_tx Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 07/20] wifi: mt76: mt7925: skip a polled link without a station Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 08/20] wifi: mt76: mt7925: skip a link without a link conf in the MLO PS update Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 09/20] wifi: mt76: mt7925: restore the beacon filter when monitor mode ends Felix Fietkau
2026-11-20 17:44 ` Felix Fietkau [this message]
2026-11-20 17:44 ` [PATCH mt76-next 11/20] wifi: mt76: mt7925: keep the default link in the active links Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 12/20] wifi: mt76: mt7925: read the monitor state under the mutex for deep sleep Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 13/20] wifi: mt76: clear roc_link before the ROC link is freed Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 14/20] wifi: mt76: mt792x: handle all interfaces of a channel context Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 15/20] wifi: mt76: mt792x: set up the sniffer when a monitor gets its channel Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 16/20] wifi: mt76: buffer Disassociation frames for stations in power save Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 17/20] wifi: mt76: mt7921: read the monitor state under the mutex for deep sleep Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 18/20] wifi: mt76: mt7925: wake the device to unassign a channel context Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 19/20] wifi: mt76: mt7996: keep the firmware log settings across a reload Felix Fietkau
2026-11-20 17:44 ` [PATCH mt76-next 20/20] wifi: mt76: mt7915: " Felix Fietkau
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=20261120174459.1657382-10-nbd@nbd.name \
--to=nbd@nbd.name \
--cc=linux-wireless@vger.kernel.org \
/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