Hi Ezequiel,

On Thu, Jan 14, 2021 at 10:47:08AM -0300, Ezequiel Garcia wrote:

On Thu, 2021-01-14 at 04:27 +0200, Laurent Pinchart wrote:

quoted

On Tue, Jan 12, 2021 at 10:23:39AM -0300, Ezequiel Garcia wrote:

quoted

Most -if not all- use-cases are expected to be covered by one of:
v4l2_async_notifier_add_fwnode_subdev,
v4l2_async_notifier_add_fwnode_remote_subdev,
v4l2_async_notifier_add_i2c_subdev or
v4l2_async_notifier_add_devname_subdev.

We'd like to discourage drivers from using v4l2_async_notifier_add_subdev,
so rename it as __v4l2_async_notifier_add_subdev. This is
typically a good hint for drivers to avoid using the function.

Signed-off-by: Ezequiel Garcia <redacted>
---
 drivers/media/v4l2-core/v4l2-async.c  | 10 +++++-----
 drivers/media/v4l2-core/v4l2-fwnode.c |  2 +-
 include/media/v4l2-async.h            | 10 ++++++++--
 3 files changed, 14 insertions(+), 8 deletions(-)

diff --git a/drivers/media/v4l2-core/v4l2-async.c b/drivers/media/v4l2-core/v4l2-async.c
index b325bacddff4..e8e690280922 100644
--- a/drivers/media/v4l2-core/v4l2-async.c
+++ b/drivers/media/v4l2-core/v4l2-async.c

@@ -630,7 +630,7 @@ void v4l2_async_notifier_cleanup(struct v4l2_async_notifier *notifier)

 }
 EXPORT_SYMBOL_GPL(v4l2_async_notifier_cleanup);
 
-int v4l2_async_notifier_add_subdev(struct v4l2_async_notifier *notifier,
+int __v4l2_async_notifier_add_subdev(struct v4l2_async_notifier *notifier,
                                   struct v4l2_async_subdev *asd)
 {
        int ret;

@@ -647,7 +647,7 @@ int v4l2_async_notifier_add_subdev(struct v4l2_async_notifier *notifier,

        mutex_unlock(&list_lock);
        return ret;
 }
-EXPORT_SYMBOL_GPL(v4l2_async_notifier_add_subdev);
+EXPORT_SYMBOL_GPL(__v4l2_async_notifier_add_subdev);
 
 struct v4l2_async_subdev *
 v4l2_async_notifier_add_fwnode_subdev(struct v4l2_async_notifier *notifier,

@@ -664,7 +664,7 @@ v4l2_async_notifier_add_fwnode_subdev(struct v4l2_async_notifier *notifier,

        asd->match_type = V4L2_ASYNC_MATCH_FWNODE;
        asd->match.fwnode = fwnode_handle_get(fwnode);
 
-       ret = v4l2_async_notifier_add_subdev(notifier, asd);
+       ret = __v4l2_async_notifier_add_subdev(notifier, asd);
        if (ret) {
                fwnode_handle_put(fwnode);
                kfree(asd);

@@ -714,7 +714,7 @@ v4l2_async_notifier_add_i2c_subdev(struct v4l2_async_notifier *notifier,

        asd->match.i2c.adapter_id = adapter_id;
        asd->match.i2c.address = address;
 
-       ret = v4l2_async_notifier_add_subdev(notifier, asd);
+       ret = __v4l2_async_notifier_add_subdev(notifier, asd);
        if (ret) {
                kfree(asd);
                return ERR_PTR(ret);

@@ -739,7 +739,7 @@ v4l2_async_notifier_add_devname_subdev(struct v4l2_async_notifier *notifier,

        asd->match_type = V4L2_ASYNC_MATCH_DEVNAME;
        asd->match.device_name = device_name;
 
-       ret = v4l2_async_notifier_add_subdev(notifier, asd);
+       ret = __v4l2_async_notifier_add_subdev(notifier, asd);
        if (ret) {
                kfree(asd);
                return ERR_PTR(ret);

diff --git a/drivers/media/v4l2-core/v4l2-fwnode.c b/drivers/media/v4l2-core/v4l2-fwnode.c
index 5353e37eb950..919fde20032e 100644
--- a/drivers/media/v4l2-core/v4l2-fwnode.c
+++ b/drivers/media/v4l2-core/v4l2-fwnode.c

@@ -833,7 +833,7 @@ v4l2_async_notifier_fwnode_parse_endpoint(struct device *dev,

        if (ret < 0)
                goto out_err;
 
-       ret = v4l2_async_notifier_add_subdev(notifier, asd);
+       ret = __v4l2_async_notifier_add_subdev(notifier, asd);
        if (ret < 0) {
                /* not an error if asd already exists */
                if (ret == -EEXIST)

I wonder if v4l2-fwnode should be moved to videodev-objs in the
Makefile, we could then avoid exporting
__v4l2_async_notifier_add_subdev(). This doesn't need to be fixed here,

Yeah, I thought about that, precisely to avoid the export and disallow
any future users of v4l2_async_notifier_add_subdev.

The impact of that is that we'd kill V4L2_FWNODE, increasing the
size of videodev.ko, by a small fraction.

We don't have to drop V4L2_FWNODE, we can still make v4l2-async.o
conditional, even if it's included in videodev.ko.

Perhaps this is justified? I'm sure all the use-cases for this type
of hardware won't care for 20 KiB more or less in the kernel payload.

quoted

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

quoted

diff --git a/include/media/v4l2-async.h b/include/media/v4l2-async.h
index 1278f98355a7..9cf83f1ecca6 100644
--- a/include/media/v4l2-async.h
+++ b/include/media/v4l2-async.h

@@ -161,17 +161,23 @@ void v4l2_async_debug_init(struct dentry *debugfs_dir);

 void v4l2_async_notifier_init(struct v4l2_async_notifier *notifier);
 
 /**
- * v4l2_async_notifier_add_subdev - Add an async subdev to the
+ * __v4l2_async_notifier_add_subdev - Add an async subdev to the
  *                             notifier's master asd list.
  *
  * @notifier: pointer to &struct v4l2_async_notifier
  * @asd: pointer to &struct v4l2_async_subdev
  *
+ * \warning: Drivers should avoid using this function and instead use one of:
+ * @v4l2_async_notifier_add_fwnode_subdev,
+ * @v4l2_async_notifier_add_fwnode_remote_subdev,
+ * @v4l2_async_notifier_add_i2c_subdev or
+ * @v4l2_async_notifier_add_devname_subdev.
+ *
  * Call this function before registering a notifier to link the provided @asd to
  * the notifiers master @asd_list. The @asd must be allocated with k*alloc() as
  * it will be freed by the framework when the notifier is destroyed.
  */
-int v4l2_async_notifier_add_subdev(struct v4l2_async_notifier *notifier,
+int __v4l2_async_notifier_add_subdev(struct v4l2_async_notifier *notifier,
                                   struct v4l2_async_subdev *asd);
 
 /**

-- 
Regards,

Laurent Pinchart