[PATCH 0/5] extcon: cancel wq_detcable delayed work on driver unbind

mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed

* [PATCH 0/5] extcon: cancel wq_detcable delayed work on driver unbind
@ 2026-10-20 15:41 Tristan Madani
  2026-10-20 15:41 ` [PATCH 1/5] extcon: max14577: cancel wq_detcable " Tristan Madani
                   ` (4 more replies)
  0 siblings, 5 replies; 6+ messages in thread
From: Tristan Madani @ 2026-10-20 15:41 UTC (permalink / raw)
  To: Chanwoo Choi, MyungJoo Ham, Krzysztof Kozlowski
  Cc: Kyungmin Park, Seung-Woo Kim, linux-kernel, stable, Tristan Madani

From: Tristan Madani <tristan@talencesecurity.com>

Several extcon MUIC drivers schedule a delayed work (wq_detcable) during
probe for post-boot cable state detection, but never cancel it on driver
unbind.  If the device is removed before the work fires, the callback
dereferences freed memory.

Use devm_delayed_work_autocancel() to tie the work lifecycle to the
device.  This matches the pattern already used by extcon-gpio,
extcon-palmas, extcon-qcom-spmi-misc, and extcon-intel-int3496.

For the Maxim drivers (max14577, max77693, max8997), irq_work was
already converted to devm_work_autocancel() but wq_detcable was missed.
The rt8973a and sm5502 drivers also need the devm-helpers.h include.

Tristan Madani (5):
  extcon: max14577: cancel wq_detcable on driver unbind
  extcon: max77693: cancel wq_detcable on driver unbind
  extcon: max8997: cancel wq_detcable on driver unbind
  extcon: rt8973a: cancel wq_detcable on driver unbind
  extcon: sm5502: cancel wq_detcable on driver unbind

 drivers/extcon/extcon-max14577.c | 5 ++++-
 drivers/extcon/extcon-max77693.c | 5 ++++-
 drivers/extcon/extcon-max8997.c  | 5 ++++-
 drivers/extcon/extcon-rt8973a.c  | 6 +++++-
 drivers/extcon/extcon-sm5502.c   | 6 +++++-
 5 files changed, 22 insertions(+), 5 deletions(-)

-- 
2.47.3

^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH 1/5] extcon: max14577: cancel wq_detcable on driver unbind
  2026-10-20 15:41 [PATCH 0/5] extcon: cancel wq_detcable delayed work on driver unbind Tristan Madani
@ 2026-10-20 15:41 ` Tristan Madani
  2026-10-20 15:41 ` [PATCH 2/5] extcon: max77693: " Tristan Madani
                   ` (3 subsequent siblings)
  4 siblings, 0 replies; 6+ messages in thread
From: Tristan Madani @ 2026-10-20 15:41 UTC (permalink / raw)
  To: Chanwoo Choi, MyungJoo Ham, Krzysztof Kozlowski
  Cc: Kyungmin Park, Seung-Woo Kim, linux-kernel, stable, Tristan Madani

From: Tristan Madani <tristan@talencesecurity.com>

The driver schedules wq_detcable as a delayed work during probe but
never cancels it.  If the device is unbound before the work fires, the
callback dereferences freed memory because the driver data is managed
by devm.

Use devm_delayed_work_autocancel() so the work is automatically
cancelled when the device is removed, matching the existing irq_work
conversion done in the same driver.

Fixes: 962e56bfcf0b ("extcon: max14577: Add extcon-max14577 driver to support MUIC device")
Cc: stable@vger.kernel.org
Signed-off-by: Tristan Madani <tristan@talencesecurity.com>
---
 drivers/extcon/extcon-max14577.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/extcon/extcon-max14577.c b/drivers/extcon/extcon-max14577.c
index 5476f48ed74b9..905722ede7ab1 100644
--- a/drivers/extcon/extcon-max14577.c
+++ b/drivers/extcon/extcon-max14577.c
@@ -763,7 +763,10 @@ static int max14577_muic_probe(struct platform_device *pdev)
 	 * After completing the booting of platform, the extcon provider
 	 * driver should notify cable state to upper layer.
 	 */
-	INIT_DELAYED_WORK(&info->wq_detcable, max14577_muic_detect_cable_wq);
+	ret = devm_delayed_work_autocancel(&pdev->dev, &info->wq_detcable,
+					   max14577_muic_detect_cable_wq);
+	if (ret)
+		return ret;
 	queue_delayed_work(system_power_efficient_wq, &info->wq_detcable,
 			delay_jiffies);
 
-- 
2.47.3


^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH 2/5] extcon: max77693: cancel wq_detcable on driver unbind
  2026-10-20 15:41 [PATCH 0/5] extcon: cancel wq_detcable delayed work on driver unbind Tristan Madani
  2026-10-20 15:41 ` [PATCH 1/5] extcon: max14577: cancel wq_detcable " Tristan Madani
@ 2026-10-20 15:41 ` Tristan Madani
  2026-10-20 15:41 ` [PATCH 3/5] extcon: max8997: " Tristan Madani
                   ` (2 subsequent siblings)
  4 siblings, 0 replies; 6+ messages in thread
From: Tristan Madani @ 2026-10-20 15:41 UTC (permalink / raw)
  To: Chanwoo Choi, MyungJoo Ham, Krzysztof Kozlowski
  Cc: Kyungmin Park, Seung-Woo Kim, linux-kernel, stable, Tristan Madani

From: Tristan Madani <tristan@talencesecurity.com>

The driver schedules wq_detcable as a delayed work during probe but
never cancels it.  If the device is unbound before the work fires, the
callback dereferences freed memory because the driver data is managed
by devm.

Use devm_delayed_work_autocancel() so the work is automatically
cancelled when the device is removed, matching the existing irq_work
conversion done in the same driver.

Fixes: 297620fd1e14 ("extcon: max77693: Check the state/type of cable after boot completed")
Cc: stable@vger.kernel.org
Signed-off-by: Tristan Madani <tristan@talencesecurity.com>
---
 drivers/extcon/extcon-max77693.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/extcon/extcon-max77693.c b/drivers/extcon/extcon-max77693.c
index 2c567e0b7b7f0..a76084e4d5512 100644
--- a/drivers/extcon/extcon-max77693.c
+++ b/drivers/extcon/extcon-max77693.c
@@ -1251,7 +1251,10 @@ static int max77693_muic_probe(struct platform_device *pdev)
 	 * After completing the booting of platform, the extcon provider
 	 * driver should notify cable state to upper layer.
 	 */
-	INIT_DELAYED_WORK(&info->wq_detcable, max77693_muic_detect_cable_wq);
+	ret = devm_delayed_work_autocancel(&pdev->dev, &info->wq_detcable,
+					   max77693_muic_detect_cable_wq);
+	if (ret)
+		return ret;
 	queue_delayed_work(system_power_efficient_wq, &info->wq_detcable,
 			delay_jiffies);
 
-- 
2.47.3


^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH 3/5] extcon: max8997: cancel wq_detcable on driver unbind
  2026-10-20 15:41 [PATCH 0/5] extcon: cancel wq_detcable delayed work on driver unbind Tristan Madani
  2026-10-20 15:41 ` [PATCH 1/5] extcon: max14577: cancel wq_detcable " Tristan Madani
  2026-10-20 15:41 ` [PATCH 2/5] extcon: max77693: " Tristan Madani
@ 2026-10-20 15:41 ` Tristan Madani
  2026-10-20 15:41 ` [PATCH 4/5] extcon: rt8973a: " Tristan Madani
  2026-10-20 15:41 ` [PATCH 5/5] extcon: sm5502: " Tristan Madani
  4 siblings, 0 replies; 6+ messages in thread
From: Tristan Madani @ 2026-10-20 15:41 UTC (permalink / raw)
  To: Chanwoo Choi, MyungJoo Ham, Krzysztof Kozlowski
  Cc: Kyungmin Park, Seung-Woo Kim, linux-kernel, stable, Tristan Madani

From: Tristan Madani <tristan@talencesecurity.com>

The driver schedules wq_detcable as a delayed work during probe but
never cancels it.  If the device is unbound before the work fires, the
callback dereferences freed memory because the driver data is managed
by devm.

Use devm_delayed_work_autocancel() so the work is automatically
cancelled when the device is removed, matching the existing irq_work
conversion done in the same driver.

Fixes: af5eb1a13273 ("extcon: max8997: Use workqueue to check cable state after completing boot of platform")
Cc: stable@vger.kernel.org
Signed-off-by: Tristan Madani <tristan@talencesecurity.com>
---
 drivers/extcon/extcon-max8997.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/extcon/extcon-max8997.c b/drivers/extcon/extcon-max8997.c
index 9cddf08e06966..7048b9f6d8443 100644
--- a/drivers/extcon/extcon-max8997.c
+++ b/drivers/extcon/extcon-max8997.c
@@ -754,7 +754,10 @@ static int max8997_muic_probe(struct platform_device *pdev)
 	 * After completing the booting of platform, the extcon provider
 	 * driver should notify cable state to upper layer.
 	 */
-	INIT_DELAYED_WORK(&info->wq_detcable, max8997_muic_detect_cable_wq);
+	ret = devm_delayed_work_autocancel(&pdev->dev, &info->wq_detcable,
+					   max8997_muic_detect_cable_wq);
+	if (ret)
+		return ret;
 	queue_delayed_work(system_power_efficient_wq, &info->wq_detcable,
 			delay_jiffies);
 
-- 
2.47.3


^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH 4/5] extcon: rt8973a: cancel wq_detcable on driver unbind
  2026-10-20 15:41 [PATCH 0/5] extcon: cancel wq_detcable delayed work on driver unbind Tristan Madani
                   ` (2 preceding siblings ...)
  2026-10-20 15:41 ` [PATCH 3/5] extcon: max8997: " Tristan Madani
@ 2026-10-20 15:41 ` Tristan Madani
  2026-10-20 15:41 ` [PATCH 5/5] extcon: sm5502: " Tristan Madani
  4 siblings, 0 replies; 6+ messages in thread
From: Tristan Madani @ 2026-10-20 15:41 UTC (permalink / raw)
  To: Chanwoo Choi, MyungJoo Ham, Krzysztof Kozlowski
  Cc: Kyungmin Park, Seung-Woo Kim, linux-kernel, stable, Tristan Madani

From: Tristan Madani <tristan@talencesecurity.com>

The driver schedules wq_detcable as a delayed work during probe but
the remove callback only tears down the regmap IRQ chip and never
cancels it.  If the device is unbound before the work fires, the
callback dereferences freed memory because the driver data is managed
by devm.

Use devm_delayed_work_autocancel() so the work is automatically
cancelled when the device is removed.

Fixes: c03e017c4d3d ("extcon: rt8973a: Add Richtek RT8973A extcon driver")
Cc: stable@vger.kernel.org
Signed-off-by: Tristan Madani <tristan@talencesecurity.com>
---
 drivers/extcon/extcon-rt8973a.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/extcon/extcon-rt8973a.c b/drivers/extcon/extcon-rt8973a.c
index 19bb49f13fb08..184e9f4c16ce3 100644
--- a/drivers/extcon/extcon-rt8973a.c
+++ b/drivers/extcon/extcon-rt8973a.c
@@ -6,6 +6,7 @@
  * Author: Chanwoo Choi <cw00.choi@samsung.com>
  */
 
+#include <linux/devm-helpers.h>
 #include <linux/err.h>
 #include <linux/i2c.h>
 #include <linux/input.h>
@@ -635,7 +636,10 @@ static int rt8973a_muic_i2c_probe(struct i2c_client *i2c)
 	 * After completing the booting of platform, the extcon provider
 	 * driver should notify cable state to upper layer.
 	 */
-	INIT_DELAYED_WORK(&info->wq_detcable, rt8973a_muic_detect_cable_wq);
+	ret = devm_delayed_work_autocancel(&i2c->dev, &info->wq_detcable,
+					   rt8973a_muic_detect_cable_wq);
+	if (ret)
+		return ret;
 	queue_delayed_work(system_power_efficient_wq, &info->wq_detcable,
 			msecs_to_jiffies(DELAY_MS_DEFAULT));
 
-- 
2.47.3


^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH 5/5] extcon: sm5502: cancel wq_detcable on driver unbind
  2026-10-20 15:41 [PATCH 0/5] extcon: cancel wq_detcable delayed work on driver unbind Tristan Madani
                   ` (3 preceding siblings ...)
  2026-10-20 15:41 ` [PATCH 4/5] extcon: rt8973a: " Tristan Madani
@ 2026-10-20 15:41 ` Tristan Madani
  4 siblings, 0 replies; 6+ messages in thread
From: Tristan Madani @ 2026-10-20 15:41 UTC (permalink / raw)
  To: Chanwoo Choi, MyungJoo Ham, Krzysztof Kozlowski
  Cc: Kyungmin Park, Seung-Woo Kim, linux-kernel, stable, Tristan Madani

From: Tristan Madani <tristan@talencesecurity.com>

sm5502_muic_i2c_probe() schedules a delayed work (wq_detcable) for
post-boot cable state detection but the driver has no .remove callback
and never cancels it.  If the device is unbound before the work fires,
the callback dereferences freed memory.

Use devm_delayed_work_autocancel() so the work is automatically
cancelled when the device is removed.

Fixes: e1954452f500 ("extcon: sm5502: Detect cable state after completing platform booting")
Cc: stable@vger.kernel.org
Signed-off-by: Tristan Madani <tristan@talencesecurity.com>
---
 drivers/extcon/extcon-sm5502.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/extcon/extcon-sm5502.c b/drivers/extcon/extcon-sm5502.c
index c8c4b9ef72aa1..5257cb3ca2266 100644
--- a/drivers/extcon/extcon-sm5502.c
+++ b/drivers/extcon/extcon-sm5502.c
@@ -6,6 +6,7 @@
  * Author: Chanwoo Choi <cw00.choi@samsung.com>
  */
 
+#include <linux/devm-helpers.h>
 #include <linux/err.h>
 #include <linux/i2c.h>
 #include <linux/interrupt.h>
@@ -763,7 +764,10 @@ static int sm5022_muic_i2c_probe(struct i2c_client *i2c)
 	 * After completing the booting of platform, the extcon provider
 	 * driver should notify cable state to upper layer.
 	 */
-	INIT_DELAYED_WORK(&info->wq_detcable, sm5502_muic_detect_cable_wq);
+	ret = devm_delayed_work_autocancel(&i2c->dev, &info->wq_detcable,
+					   sm5502_muic_detect_cable_wq);
+	if (ret)
+		return ret;
 	queue_delayed_work(system_power_efficient_wq, &info->wq_detcable,
 			msecs_to_jiffies(DELAY_MS_DEFAULT));
 
-- 
2.47.3


^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2026-10-20 15:41 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-10-20 15:41 [PATCH 0/5] extcon: cancel wq_detcable delayed work on driver unbind Tristan Madani
2026-10-20 15:41 ` [PATCH 1/5] extcon: max14577: cancel wq_detcable " Tristan Madani
2026-10-20 15:41 ` [PATCH 2/5] extcon: max77693: " Tristan Madani
2026-10-20 15:41 ` [PATCH 3/5] extcon: max8997: " Tristan Madani
2026-10-20 15:41 ` [PATCH 4/5] extcon: rt8973a: " Tristan Madani
2026-10-20 15:41 ` [PATCH 5/5] extcon: sm5502: " Tristan Madani

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®