Skip to contentSkip to conversation

staging: most: video: replace BUG_ON() with pr_warn() in comp_exit()

PATCH In review Gabriel Rondon 2 participants 4 messages Apr 8 – Apr 12

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 staging: most: video: replace BUG_ON() with pr_warn() in comp_exit()

Replace BUG_ON(!list_empty(&video_devices)) with an explicit check and pr_warn() in the module exit function.

BUG_ON() is deprecated as it crashes the entire kernel on assertion failure (see Documentation/process/deprecated.rst). WARN_ON() is also inappropriate here since it crashes the system when panic-on-warn is enabled. Use pr_warn() to log the unexpected condition without any risk of crashing.

This is the last remaining BUG_ON() in the staging/most subsystem, following the dim2 series that replaced five BUG_ON() calls.

---
Changes since v1:
- Use pr_warn() instead of WARN_ON() to avoid crash with panic-on-warn (Greg Kroah-Hartman)

 drivers/staging/most/video/video.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/staging/most/video/video.c +2 -1
index 04351f8ccccf..c5425c207552 100644
@@ -577,7 +577,8 @@ static void __exit comp_exit(void)
 
 	most_deregister_configfs_subsys(&comp);
 	most_deregister_component(&comp);
-	BUG_ON(!list_empty(&video_devices));
+	if (!list_empty(&video_devices))
+		pr_warn("video_devices list not empty on exit\n");
 }
 
 module_init(comp_init);
Signature · 1 line
2.33.0
#1 lore raw
Quoted 26 lines · Gabriel Rondon

Replace BUG_ON(!list_empty(&video_devices)) with an explicit check and pr_warn() in the module exit function.

BUG_ON() is deprecated as it crashes the entire kernel on assertion failure (see Documentation/process/deprecated.rst). WARN_ON() is also inappropriate here since it crashes the system when panic-on-warn is enabled. Use pr_warn() to log the unexpected condition without any risk of crashing.

This is the last remaining BUG_ON() in the staging/most subsystem, following the dim2 series that replaced five BUG_ON() calls.

---
Changes since v1:
- Use pr_warn() instead of WARN_ON() to avoid crash with panic-on-warn (Greg Kroah-Hartman)

 drivers/staging/most/video/video.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/staging/most/video/video.c +0 -0
index 04351f8ccccf..c5425c207552 100644

drivers/staging/most/video/video.c static void __exit comp_exit(void) · line 581

Quoted 6 lines
drivers/staging/most/video/video.c +2 -1
@@ -577,7 +577,8 @@ static void __exit comp_exit(void)
 
 	most_deregister_configfs_subsys(&comp);
 	most_deregister_component(&comp);
-	BUG_ON(!list_empty(&video_devices));
+	if (!list_empty(&video_devices))
+		pr_warn("video_devices list not empty on exit\n");

How can this ever happen? Why not fix that to ensure it can never happen, and then this check can just be removed entirely?

thanks,

greg k-h

#2 lore raw
Quoted 2 lines · Greg Kroah-Hartman

How can this ever happen? Why not fix that to ensure it can never happen, and then this check can just be removed entirely?

Thanks for the review, Greg.

I looked at most_deregister_component() in core.c — it already calls disconnect_channel() for every linked channel via bus_for_each_dev(). So the manual cleanup loop in comp_exit() and the comment saying "mostcore currently doesn't call disconnect_channel()" are outdated.

For v3 I plan to remove the manual cleanup loop, the stale comment, and the BUG_ON entirely — leaving comp_exit() to just call most_deregister_configfs_subsys() and most_deregister_component(), which already handles the full teardown.

Does that match what you had in mind?

Gabriel

#3 lore raw

PATCH v3 1/1 staging: most: video: remove redundant cleanup in comp_exit()

In reply to Greg Kroah-Hartman

most_deregister_component() already calls disconnect_channel() for every linked channel via bus_for_each_dev() in core.c, which invokes comp_disconnect_channel() to remove each entry from the video_devices list and tear down the V4L2 device.

The manual cleanup loop in comp_exit() duplicates this work and is guarded by a stale comment claiming that "mostcore currently doesn't call disconnect_channel() for linked channels" — but the core has since been fixed to do exactly that.

Remove the redundant manual cleanup loop, the outdated comment, and the BUG_ON() assertion that checked for a condition that can no longer occur.

---
Changes since v2:
- Remove the cleanup loop and BUG_ON() entirely instead of replacing BUG_ON() with pr_warn(). most_deregister_component() already calls disconnect_channel() for every linked channel via bus_for_each_dev() in core.c, which handles the teardown. (Greg Kroah-Hartman)

Changes since v1:
- v2 used pr_warn() instead of WARN_ON() to avoid a crash with panic-on-warn (Greg Kroah-Hartman). Superseded by the v3 approach above.

 drivers/staging/most/video/video.c | 21 ---------------------
 1 file changed, 21 deletions(-)
drivers/staging/most/video/video.c +0 -21
index 04351f8ccccf..98988b0d9330 100644
@@ -555,29 +555,8 @@ static int __init comp_init(void)
 
 static void __exit comp_exit(void)
 {
-	struct most_video_dev *mdev, *tmp;
-	LIST_HEAD(free_list);
-
-	/*
-	 * As the mostcore currently doesn't call disconnect_channel()
-	 * for linked channels while we call most_deregister_component()
-	 * we simulate this call here.
-	 * This must be fixed in core.
-	 */
-	spin_lock_irq(&list_lock);
-	list_replace_init(&video_devices, &free_list);
-	spin_unlock_irq(&list_lock);
-
-	list_for_each_entry_safe(mdev, tmp, &free_list, list) {
-		list_del_init(&mdev->list);
-		comp_unregister_videodev(mdev);
-		v4l2_device_disconnect(&mdev->v4l2_dev);
-		v4l2_device_put(&mdev->v4l2_dev);
-	}
-
 	most_deregister_configfs_subsys(&comp);
 	most_deregister_component(&comp);
-	BUG_ON(!list_empty(&video_devices));
 }
 
 module_init(comp_init);
Signature · 1 line
2.33.0