Linux-mediatek Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Sean Wang <sean.wang@kernel.org>
To: nbd@nbd.name, lorenzo.bianconi@kernel.org
Cc: yu-ching.liu@mediatek.com, chengwei.yu@mediatek.com,
	jacobs.wu@mediatek.com, posh.sun@mediatek.com,
	clifford.hsu@mediatek.com, linux-wireless@vger.kernel.org,
	linux-mediatek@lists.infradead.org,
	Sean Wang <sean.wang@mediatek.com>
Subject: [PATCH v2 16/19] wifi: mt76: mt7925: stop queueing resets once the device is being removed
Date: Mon, 16 Nov 2026 17:16:27 -0600	[thread overview]
Message-ID: <20261116231630.1733906-17-sean.wang@kernel.org> (raw)
In-Reply-To: <20261116231630.1733906-1-sean.wang@kernel.org>

From: Jacobs Wu <jacobs.wu@mediatek.com>

mt7925e_unregister_device() cancels reset_work before it tears the device
down, but every source that can raise a reset stays live past that point:
the MCU command path, the system error recovery and interrupt handlers,
and the MAC watchdog all
call mt792x_reset(), and the interrupt tasklet is only disabled at the very
end of the function. A reset raised in that window is queued behind the
cancel and then runs while the device is being dismantled -
mt7925_mac_reset_work() sets hw_full_reset, stops the queues and
cancels the PM works before it can notice that the device is gone.

mt792x_reset() already bails out on !hw_init_done, and that flag has no
other consumer: it is set once during hardware init and read only there.
Clear it at the top of the teardown so no reset can be queued for its whole
duration.

Measured on rauru with kprobes on mt792x_reset() (queue), on
mt7925_mac_reset_work() (execution) and on mt76_unregister_device(), which
runs immediately after the cancel and so serves as the anchor, while
chip_reset was written in a loop across the module unload:

  before: 205 resets queued and 415 mt7925_mac_reset_work() runs after the
          anchor, that is after cancel_work_sync() had already returned
  after:  no run after the anchor; mt792x_reset() is still entered but
          returns early

A chip_reset on a running device still triggers a reset as before, and
unload/reload cycles stay clean.

Fixes: c948b5da6bbe ("wifi: mt76: mt7925: add Mediatek Wi-Fi7 driver for mt7925 chips")
Co-developed-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Jacobs Wu <jacobs.wu@mediatek.com>
---
 drivers/net/wireless/mediatek/mt76/mt7925/pci.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/drivers/net/wireless/mediatek/mt76/mt7925/pci.c b/drivers/net/wireless/mediatek/mt76/mt7925/pci.c
index 8d0f8f72d2f9..fe168e00b9eb 100644
--- a/drivers/net/wireless/mediatek/mt76/mt7925/pci.c
+++ b/drivers/net/wireless/mediatek/mt76/mt7925/pci.c
@@ -103,6 +103,16 @@ static void mt7925e_unregister_device(struct mt792x_dev *dev)
 	if (dev->phy.chip_cap & MT792x_CHIP_CAP_WF_RF_PIN_CTRL_EVT_EN)
 		wiphy_rfkill_stop_polling(hw->wiphy);
 
+	/* Stop new resets from being queued for the rest of the teardown.
+	 * mt792x_reset() bails out on !hw_init_done, which is otherwise only
+	 * set once at init, so clearing it here closes the window in which an
+	 * MCU timeout, a system error recovery interrupt or the watchdog
+	 * could still schedule
+	 * reset_work behind the cancel below and run it against a device that
+	 * is already being dismantled.
+	 */
+	dev->hw_init_done = false;
+
 	cancel_work_sync(&dev->init_work);
 	mt76_unregister_device(&dev->mt76);
 	set_bit(MT76_REMOVED, &dev->mphy.state);
-- 
2.43.0



  parent reply	other threads:[~2026-11-16 23:20 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-16 23:16 [PATCH v2 00/19] wifi: mt76: mt7925: add NAN security and improve NDP reliability (part 2) Sean Wang
2026-11-16 23:16 ` [PATCH v2 01/19] wifi: mt76: mt7925: add NAN low power event control on suspend/resume Sean Wang
2026-11-16 23:16 ` [PATCH v2 02/19] wifi: mt76: mt7925: implement NAN key management MCU command Sean Wang
2026-11-16 23:16 ` [PATCH v2 03/19] wifi: mt76: mt7925: add per-peer RX IGTK/BIGTK and RX GTK WTBLs Sean Wang
2026-11-16 23:16 ` [PATCH v2 04/19] wifi: mt76: mt7925: implement mt7925_nan_set_key for NAN security Sean Wang
2026-11-16 23:16 ` [PATCH v2 05/19] wifi: mt76: mt7925: install NAN BIP keys standalone in sta_key_tlv Sean Wang
2026-11-16 23:16 ` [PATCH v2 06/19] wifi: mt76: mt7925: do not disable RX NAPI twice on unload Sean Wang
2026-11-16 23:16 ` [PATCH v2 07/19] wifi: mt76: mt7925: assign the interface WTBL to the NAN management TXQ Sean Wang
2026-11-16 23:16 ` [PATCH v2 08/19] wifi: mt76: mt7925: do not deactivate shared NAN BSS on peer STA removal Sean Wang
2026-11-16 23:16 ` [PATCH v2 09/19] wifi: mt76: gate NAN unicast mgmt by peer availability via STA queues Sean Wang
2026-11-16 23:16 ` [PATCH v2 10/19] wifi: mt76: mt7925: hold NAN mgmt TX to the discovery window Sean Wang
2026-11-16 23:16 ` [PATCH v2 11/19] wifi: mt76: mt7925: double the retry budget for NAN unicast management Sean Wang
2026-11-16 23:16 ` [PATCH v2 12/19] wifi: mt76: mt7925: let TX status of DW-held NAN frames outlive the DW period Sean Wang
2026-11-16 23:16 ` [PATCH v2 13/19] wifi: mt76: mt7925: defer the NAN joined-cluster event out of NAN_START Sean Wang
2026-11-16 23:16 ` [PATCH v2 14/19] wifi: mt76: mt7925: steer NAN handshake frames by committed-bitmap state Sean Wang
2026-11-16 23:16 ` [PATCH v2 15/19] wifi: mt76: mt7925: disable only the RX NAPI instances that exist Sean Wang
2026-11-16 23:16 ` Sean Wang [this message]
2026-11-16 23:16 ` [PATCH v2 17/19] wifi: mt76: mt7925: bound the lifetime of NAN unicast management frames Sean Wang
2026-11-16 23:16 ` [PATCH v2 18/19] wifi: mt76: mt7925: always deliver the joined-cluster event through the deferred work Sean Wang
2026-11-16 23:16 ` [PATCH v2 19/19] wifi: mt76: mt792x: allow three channels in the NAN combination Sean Wang

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=20261116231630.1733906-17-sean.wang@kernel.org \
    --to=sean.wang@kernel.org \
    --cc=chengwei.yu@mediatek.com \
    --cc=clifford.hsu@mediatek.com \
    --cc=jacobs.wu@mediatek.com \
    --cc=linux-mediatek@lists.infradead.org \
    --cc=linux-wireless@vger.kernel.org \
    --cc=lorenzo.bianconi@kernel.org \
    --cc=nbd@nbd.name \
    --cc=posh.sun@mediatek.com \
    --cc=sean.wang@mediatek.com \
    --cc=yu-ching.liu@mediatek.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