Re: [alsa-devel] [PATCH v4 07/15] soundwire: slave: move uevent handling to slave

[Pierre-Louis Bossart]

On 12/13/19 1:22 AM, Greg KH wrote:
> On Thu, Dec 12, 2019 at 11:04:01PM -0600, Pierre-Louis Bossart wrote:
> > Currently the code deals with uevents at the bus level, but we only care
> > for Slave events
>
> What does this mean?  I can't understand it, can you please provide more
> information on what you are doing here?

In the earlier versions of the patch, the code looks like this and there 
was an open on what to do with a master-specific event.

 static int sdw_uevent(struct device *dev, struct kobj_uevent_env *env)
 {
+	struct sdw_master_device *md;
 	struct sdw_slave *slave;
 	char modalias[32];

-	if (is_sdw_slave(dev)) {
+	if (is_sdw_md(dev)) {
+		md = to_sdw_master_device(dev);
+		/* TODO: do we need to call add_uevent_var() ? */
+	} else if (is_sdw_slave(dev)) {
 		slave = to_sdw_slave_device(dev);
+
+		sdw_slave_modalias(slave, modalias, sizeof(modalias));
+
+		if (add_uevent_var(env, "MODALIAS=%s", modalias))
+			return -ENOMEM;
 	} else {
 		dev_warn(dev, "uevent for unknown Soundwire type\n");
 		return -EINVAL;
 	}

Vinod suggested this was not needed and suggested the code for uevents 
be moved to be slave-specific, which is what this patch does.
> > diff --git a/drivers/soundwire/slave.c b/drivers/soundwire/slave.c
> > index c87267f12a3b..014c3ece1f17 100644
> > --- a/drivers/soundwire/slave.c
> > +++ b/drivers/soundwire/slave.c
> > @@ -17,6 +17,7 @@ static void sdw_slave_release(struct device *dev)
> >   struct device_type sdw_slave_type = {
> >   	.name =		"sdw_slave",
> >   	.release =	sdw_slave_release,
> > +	.uevent = sdw_uevent,
>
> Align this with the other ones?
>
> does this cause any different functionality?

As mentioned above, this move was suggested by Vinod. I don't have a 
specific need for uevents for the master and there's no functionality 
limitation, that said this is way beyond my comfort zone so I will 
follow recommendations, if any.