[PATCH] android: binderfs: drop manual unlock via scoped_guard()

[PATCH] android: binderfs: drop manual unlock via scoped_guard()

[Biren Pandya]

The device_create function manually protects the minor number allocation
with binderfs_minors_mutex and relies on multiple manual unlocks on
success and error paths. This is error-prone and adds boilerplate.

Refactor the critical section to use the modern scoped_guard(mutex)
macro. This strictly binds the lock lifecycle to the scope block,
completely eliminating the manual unlocks and guaranteeing lock
safety against future code modifications.

Signed-off-by: Biren Pandya <birenpandya@gmail.com>
---
 drivers/android/binderfs.c | 25 ++++++++++++-------------
 1 file changed, 12 insertions(+), 13 deletions(-)

diff --git a/drivers/android/binderfs.c b/drivers/android/binderfs.c
index 361d69f756f5..ad5d9439ff87 100644
--- a/drivers/android/binderfs.c
+++ b/drivers/android/binderfs.c
@@ -129,20 +129,19 @@ static int binderfs_binder_device_create(struct inode *ref_inode,
 #endif
 
 	/* Reserve new minor number for the new device. */
-	mutex_lock(&binderfs_minors_mutex);
-	if (++info->device_count <= info->mount_opts.max)
-		minor = ida_alloc_max(&binderfs_minors,
-				      use_reserve ? BINDERFS_MAX_MINOR - 1 :
-						    BINDERFS_MAX_MINOR_CAPPED - 1,
-				      GFP_KERNEL);
-	else
-		minor = -ENOSPC;
-	if (minor < 0) {
-		--info->device_count;
-		mutex_unlock(&binderfs_minors_mutex);
-		return minor;
+	scoped_guard(mutex, &binderfs_minors_mutex) {
+		if (++info->device_count <= info->mount_opts.max)
+			minor = ida_alloc_max(&binderfs_minors,
+					      use_reserve ? BINDERFS_MAX_MINOR - 1 :
+							    BINDERFS_MAX_MINOR_CAPPED - 1,
+					      GFP_KERNEL);
+		else
+			minor = -ENOSPC;
+		if (minor < 0) {
+			--info->device_count;
+			return minor;
+		}
 	}
-	mutex_unlock(&binderfs_minors_mutex);
 
 	ret = -ENOMEM;
 	device = kzalloc_obj(*device);

base-commit: 64d712aa31f30a125291e7c47209ef7ebd3285a3
-- 
2.50.1 (Apple Git-155)

Re: [PATCH] android: binderfs: drop manual unlock via scoped_guard()

[Greg KH]

On Tue, Jun 16, 2026 at 11:23:41PM +0530, Biren Pandya wrote:
> The device_create function manually protects the minor number allocation
> with binderfs_minors_mutex and relies on multiple manual unlocks on
> success and error paths. This is error-prone and adds boilerplate.
> 
> Refactor the critical section to use the modern scoped_guard(mutex)
> macro. This strictly binds the lock lifecycle to the scope block,
> completely eliminating the manual unlocks and guaranteeing lock
> safety against future code modifications.

We don't do scope guard changes that don't actually fix a bug, as it's
not needed.  Otherwise we would be rewriting large chunks of the kernel
for it.

thanks,

greg k-h