Skip to contentSkip to conversation

EDAC/device: Serialize poll_msec updates against device teardown

PATCH In review Jad Keskes 2 participants 2 messages Aug 8 – Aug 10 EDAC-core

lore
Share & tools
lore

Jad Keskes

author
#0 lore raw

PATCH 1/1 EDAC/device: Serialize poll_msec updates against device teardown

The poll_msec attribute was writable on interrupt-driven controllers, where the value is meaningless, and a write racing with device removal could re-arm the polling workqueue after it had been stopped and the ctl_info freed -- use-after-free.

Restrict the attribute to controllers that are actually polled: _show() and _store() now return -EPERM unless the controller is in the OP_RUNNING_POLL state.

That check alone is not sufficient against the teardown race: _store() can pass it and then be preempted by a concurrent del_device() that sets OP_OFFLINE, stops the workqueue and frees the ctl_info once the store completes. So re-check the state inside reset_delay_period() under device_ctls_mutex -- the same lock under which del_device() updates op_state -- making the check atomic with the re-arm. The mutex is taken only around the check and re-arm, never across the stop/teardown path, so it cannot deadlock against other pollers queued on the shared EDAC workqueue.

---
 drivers/edac/edac_device.c       | 11 +++++++++++
 drivers/edac/edac_device_sysfs.c |  8 ++++++++
 2 files changed, 19 insertions(+)
drivers/edac/edac_device.c +11 -0
index 361dc985497b..519d1c6d0930 100644
@@ -392,10 +392,21 @@ static void edac_device_workq_teardown(struct edac_device_ctl_info *edac_dev)
  */
 void edac_device_reset_delay_period(struct edac_device_ctl_info *edac_dev, unsigned int msec)
 {
+	mutex_lock(&device_ctls_mutex);
+
+	/* Don't re-arm the workqueue once teardown has begun or when the
+	 * controller isn't polled; this check must be atomic with the
+	 * re-arm below so it can't race a concurrent del_device().
+	 */
+	if (edac_dev->op_state != OP_RUNNING_POLL)
+		goto out;
+
 	edac_dev->poll_msec = msec;
 	edac_dev->delay     = msecs_to_jiffies(msec);
 
 	edac_mod_work(&edac_dev->work, edac_dev->delay);
+out:
+	mutex_unlock(&device_ctls_mutex);
 }
 
 int edac_device_alloc_index(void)
drivers/edac/edac_device_sysfs.c +8 -0
index e12122b2f42e..821a80806086 100644
@@ -83,6 +83,10 @@ static ssize_t edac_device_ctl_panic_on_ue_store(struct edac_device_ctl_info
 static ssize_t edac_device_ctl_poll_msec_show(struct edac_device_ctl_info
 					*ctl_info, char *data)
 {
+	/* Interval is only meaningful while running under polling */
+	if (ctl_info->op_state != OP_RUNNING_POLL)
+		return -EPERM;
+
 	return sprintf(data, "%u\n", ctl_info->poll_msec);
 }
 
@@ -93,6 +97,10 @@ static ssize_t edac_device_ctl_poll_msec_store(struct edac_device_ctl_info
 	unsigned int value;
 	int ret;
 
+	/* Only meaningful when the device is running under polling */
+	if (ctl_info->op_state != OP_RUNNING_POLL)
+		return -EPERM;
+
 	/*
 	 * Get the value, make sure it is non-zero, must be at least one
 	 * millisecond for the delay period between scans.
Signature · 1 line
2.55.0
maintainer
#1 lore raw
Quoted 1 line · Jad Keskes

The poll_msec attribute was writable on interrupt-driven controllers,

Controllers?

Quoted 18 lines

where the value is meaningless, and a write racing with device removal could re-arm the polling workqueue after it had been stopped and the ctl_info freed -- use-after-free.

Restrict the attribute to controllers that are actually polled: _show() and _store() now return -EPERM unless the controller is in the OP_RUNNING_POLL state.

That check alone is not sufficient against the teardown race: _store() can pass it and then be preempted by a concurrent del_device() that sets OP_OFFLINE, stops the workqueue and frees the ctl_info once the store completes. So re-check the state inside reset_delay_period() under device_ctls_mutex -- the same lock under which del_device() updates op_state -- making the check atomic with the re-arm. The mutex is taken only around the check and re-arm, never across the stop/teardown path, so it cannot deadlock against other pollers queued on the shared EDAC workqueue.

Quoted 1 line

This is not how this tag should be used. Please take the time to read up on the development process before submitting more patches:

https://www.kernel.org/doc/html/latest/process/development-process.html

This too:

https://www.kernel.org/doc/html/latest/process/submitting-patches.html

Also, how much of this patch has been done with the help of LLM?

We have a tag for that - Assisted-by:

Quoted 11 lines
---
 drivers/edac/edac_device.c       | 11 +++++++++++
 drivers/edac/edac_device_sysfs.c |  8 ++++++++
 2 files changed, 19 insertions(+)
drivers/edac/edac_device.c +0 -0
index 361dc985497b..519d1c6d0930 100644

drivers/edac/edac_device.c static void edac_device_workq_teardown(struct edac_device_ctl_info *edac_dev) · line 400

Quoted 9 lines
drivers/edac/edac_device.c +6 -0
@@ -392,10 +392,21 @@ static void edac_device_workq_teardown(struct edac_device_ctl_info *edac_dev)
  */
 void edac_device_reset_delay_period(struct edac_device_ctl_info *edac_dev, unsigned int msec)
 {
+	mutex_lock(&device_ctls_mutex);
+
+	/* Don't re-arm the workqueue once teardown has begun or when the
+	 * controller isn't polled; this check must be atomic with the
+	 * re-arm below so it can't race a concurrent del_device().
+	 */

This is not the right comment format.

Quoted 17 lines
drivers/edac/edac_device.c +5 -0
+	if (edac_dev->op_state != OP_RUNNING_POLL)
+		goto out;
+
 	edac_dev->poll_msec = msec;
 	edac_dev->delay     = msecs_to_jiffies(msec);
 
 	edac_mod_work(&edac_dev->work, edac_dev->delay);
+out:
+	mutex_unlock(&device_ctls_mutex);
 }
 
 int edac_device_alloc_index(void)
drivers/edac/edac_device_sysfs.c +0 -0
index e12122b2f42e..821a80806086 100644

drivers/edac/edac_device_sysfs.c static ssize_t edac_device_ctl_panic_on_ue_store(struct edac_device_ctl_info · line 86

Quoted 4 lines
drivers/edac/edac_device_sysfs.c +1 -0
@@ -83,6 +83,10 @@ static ssize_t edac_device_ctl_panic_on_ue_store(struct edac_device_ctl_info
 static ssize_t edac_device_ctl_poll_msec_show(struct edac_device_ctl_info
 					*ctl_info, char *data)
 {
+	/* Interval is only meaningful while running under polling */

An LLM usually slaps meaningless comments like that.

drivers/edac/edac_device_sysfs.c static ssize_t edac_device_ctl_poll_msec_store(struct edac_device_ctl_info · line 100

Quoted 11 lines
drivers/edac/edac_device_sysfs.c +4 -0
+	if (ctl_info->op_state != OP_RUNNING_POLL)
+		return -EPERM;
+
 	return sprintf(data, "%u\n", ctl_info->poll_msec);
 }
 
@@ -93,6 +97,10 @@ static ssize_t edac_device_ctl_poll_msec_store(struct edac_device_ctl_info
 	unsigned int value;
 	int ret;
 
+	/* Only meaningful when the device is running under polling */

Here too.

Quoted 9 lines
drivers/edac/edac_device_sysfs.c +3 -0
+	if (ctl_info->op_state != OP_RUNNING_POLL)
+		return -EPERM;
+
 	/*
 	 * Get the value, make sure it is non-zero, must be at least one
 	 * millisecond for the delay period between scans.
Signature · 1 line
2.55.0
Signature · 4 lines
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette