From: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>

Export the address of the /sys/class kobject to users that need to
chain off of it by means other than calling register_class(). This will
be used by the GPIO subsystem to provide a backward compatibility
mount-point for the GPIO sysfs class once it's disabled.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
---
 drivers/base/class.c         | 6 ++++++
 include/linux/device/class.h | 6 ++++++
 2 files changed, 12 insertions(+)

diff --git a/drivers/base/class.c b/drivers/base/class.c
index cb5359235c70..68474aff53df 100644
--- a/drivers/base/class.c
+++ b/drivers/base/class.c
@@ -23,6 +23,9 @@
 /* /sys/class */
 static struct kset *class_kset;
 
+struct kobject *class_kobj;
+EXPORT_SYMBOL_GPL(class_kobj);
+
 #define to_class_attr(_attr) container_of(_attr, struct class_attribute, attr)
 
 /**
@@ -668,5 +671,8 @@ int __init classes_init(void)
 	class_kset = kset_create_and_add("class", NULL, NULL);
 	if (!class_kset)
 		return -ENOMEM;
+
+	class_kobj = &class_kset->kobj;
+
 	return 0;
 }
diff --git a/include/linux/device/class.h b/include/linux/device/class.h
index 518c9c83d64b..898f7948293d 100644
--- a/include/linux/device/class.h
+++ b/include/linux/device/class.h
@@ -69,6 +69,12 @@ struct class {
 	const struct dev_pm_ops *pm;
 };
 
+/*
+ * Global /sys/class kobject for creating sub-groups by means other than
+ * class_register().
+ */
+extern struct kobject *class_kobj;
+
 struct class_dev_iter {
 	struct klist_iter		ki;
 	const struct device_type	*type;

-- 
2.43.0

On Tue, Oct 15, 2024 at 10:00:23AM +0200, Bartosz Golaszewski wrote:
> From: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
> 
> Export the address of the /sys/class kobject to users that need to
> chain off of it by means other than calling register_class().

Ick, no.

> This will
> be used by the GPIO subsystem to provide a backward compatibility
> mount-point for the GPIO sysfs class once it's disabled.

Again, ick, no.

No "mount point" should be messing with sysfs, don't do that.

sysfs is a simple "if the file or directory is not there, don't worry
about it" type of interface (i.e. fixing the issues we had with /proc).
If a userspace tool can't find something there, then it should just
error out or move on.

Let's not allow anything to mount anything at /sys/class/ please, that
way lies madness for the next 40+ years.  If you want to drop a
userspace api, drop it, don't paper over it with something that you
can't drop either.

thanks,

greg k-h