public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* PATCH: add class_device_find
@ 2004-03-02 15:49 Jamey Hicks
  2004-03-02 20:06 ` Greg KH
  0 siblings, 1 reply; 2+ messages in thread
From: Jamey Hicks @ 2004-03-02 15:49 UTC (permalink / raw)
  To: greg; +Cc: linux-kernel


This patch adds:

struct class_device * class_device_find(struct class *class, const char 
*class_id)

to find a class device by name so that drivers that match up class 
devices by ID do not need to reach into the internals of class 
implementation.  RMK recommended that I take this approach, and it seems 
reasonable to me.  Please let me know what you think.

-Jamey Hicks


--- linux-2.6.3/drivers/base/class.c    2004-03-02 10:44:56.000000000 -0500
+++ linux-2.6.3-hh1/drivers/base/class.c    2004-03-02 
10:45:14.000000000 -0500
@@ -372,6 +372,19 @@
     return 0;
 }
 
+struct class_device * class_device_find(struct class *class, const char 
*class_id)
+{
+    struct list_head * entry;
+    list_for_each(entry, &class->children) {
+        struct class_device *class_dev = container_of(entry, struct 
class_device, node);
+        if (class_dev) {
+            if (strcmp(class_dev->class_id, class_id) == 0)
+                return class_dev;
+        }
+    }
+    return NULL;
+}
+
 struct class_device * class_device_get(struct class_device *class_dev)
 {
     if (class_dev)
@@ -463,6 +476,7 @@
 EXPORT_SYMBOL(class_device_put);
 EXPORT_SYMBOL(class_device_create_file);
 EXPORT_SYMBOL(class_device_remove_file);
+EXPORT_SYMBOL(class_device_find);
 
 EXPORT_SYMBOL(class_interface_register);
 EXPORT_SYMBOL(class_interface_unregister);



^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: PATCH: add class_device_find
  2004-03-02 15:49 PATCH: add class_device_find Jamey Hicks
@ 2004-03-02 20:06 ` Greg KH
  0 siblings, 0 replies; 2+ messages in thread
From: Greg KH @ 2004-03-02 20:06 UTC (permalink / raw)
  To: Jamey Hicks; +Cc: linux-kernel

On Tue, Mar 02, 2004 at 10:49:04AM -0500, Jamey Hicks wrote:
> 
> This patch adds:
> 
> struct class_device * class_device_find(struct class *class, const char 
> *class_id)
> 
> to find a class device by name so that drivers that match up class 
> devices by ID do not need to reach into the internals of class 
> implementation.  RMK recommended that I take this approach, and it seems 
> reasonable to me.  Please let me know what you think.

Hm, well your patch was line-wrapped, the code was not race-safe, you
didn't protect the class_device that you returned, you used
list_for_each instead of list_for_each_entry, you made a check that can
never fail, and you forgot to modify device.h.

But other than that it was fine :)

How about this version instead?

Might I ask about which part of the kernel you are going to want to use
this call in?

thanks,

greg k-h


# Driver core: add class_device_find function
# 
# based on an idea from Jamey Hicks <jamey.hicks@hp.com>

diff -Nru a/drivers/base/class.c b/drivers/base/class.c
--- a/drivers/base/class.c	Tue Mar  2 12:02:30 2004
+++ b/drivers/base/class.c	Tue Mar  2 12:02:30 2004
@@ -375,6 +375,33 @@
 	return 0;
 }
 
+/**
+ * class_device_find - find a struct class_device in a specific class
+ * @class: the class to search
+ * @class_id: the class_id to search for
+ *
+ * Iterates through the list of all class devices registered to a class.  If
+ * the class_id is found, its reference count is incremented and returned to
+ * the caller.  If the class_id does not match any existing struct class_device
+ * registered to this struct class, then NULL is returned.
+ */
+struct class_device * class_device_find(struct class *class, const char *class_id)
+{
+	struct class_device *class_dev;
+	struct class_device *found = NULL;
+
+	down_read(&class->subsys.rwsem);
+	list_for_each_entry(class_dev, &class->children, node) {
+		if (strcmp(class_dev->class_id, class_id) == 0) {
+			found = class_device_get(class_dev);
+			break;
+		}
+	}
+	up_read(&class->subsys.rwsem);
+
+	return found;
+}
+
 struct class_device * class_device_get(struct class_device *class_dev)
 {
 	if (class_dev)
@@ -466,6 +493,7 @@
 EXPORT_SYMBOL(class_device_put);
 EXPORT_SYMBOL(class_device_create_file);
 EXPORT_SYMBOL(class_device_remove_file);
+EXPORT_SYMBOL(class_device_find);
 
 EXPORT_SYMBOL(class_interface_register);
 EXPORT_SYMBOL(class_interface_unregister);
diff -Nru a/include/linux/device.h b/include/linux/device.h
--- a/include/linux/device.h	Tue Mar  2 12:02:30 2004
+++ b/include/linux/device.h	Tue Mar  2 12:02:30 2004
@@ -214,6 +214,7 @@
 extern void class_device_del(struct class_device *);
 
 extern int class_device_rename(struct class_device *, char *);
+extern struct class_device * class_device_find(struct class *class, const char *class_id);
 
 extern struct class_device * class_device_get(struct class_device *);
 extern void class_device_put(struct class_device *);

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2004-03-02 20:07 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-03-02 15:49 PATCH: add class_device_find Jamey Hicks
2004-03-02 20:06 ` Greg KH

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox