Re: [PATCH 3/6] i2c: Ignore return value of i2c_del_adapter()

[Jean Delvare <khali-PUYAD+kWke1g9hUCZPvPmw@public.gmane.org>]

On Sat, 30 Mar 2013 09:11:54 +0100, Lars-Peter Clausen wrote:
> Hi Jean,
> 
> On 03/30/2013 08:55 AM, Jean Delvare wrote:
> > Hi Lars-Peter,
> > 
> > On Sat,  9 Mar 2013 19:16:46 +0100, Lars-Peter Clausen wrote:
> >> i2c_del_adapter() always returns 0.
> > 
> > I beg you pardon? i2c_del_adapter() starts with:
> > 
> > int i2c_del_adapter(struct i2c_adapter *adap)
> > {
> > 	int res = 0;
> > 	struct i2c_adapter *found;
> > 	struct i2c_client *client, *next;
> > 
> > 	/* First make sure that this adapter was ever added */
> > 	mutex_lock(&core_lock);
> > 	found = idr_find(&i2c_adapter_idr, adap->nr);
> > 	mutex_unlock(&core_lock);
> > 	if (found != adap) {
> > 		pr_debug("i2c-core: attempting to delete unregistered "
> > 			 "adapter [%s]\n", adap->name);
> > 		return -EINVAL;
> > 	}
> > 
> > 	/* Tell drivers about this removal */
> > 	mutex_lock(&core_lock);
> > 	res = bus_for_each_drv(&i2c_bus_type, NULL, adap,
> > 			       __process_removed_adapter);
> > 	mutex_unlock(&core_lock);
> > 	if (res)
> > 		return res;
> > (...)
> > 
> > So, no, it doesn't "always return 0".
> > 
> 
> Patch 1 and 2 in this series remove those two instances:
> https://lkml.org/lkml/2013/3/9/72
> https://lkml.org/lkml/2013/3/9/86
> 
> For an explanation why this should be done please take a look at the cover
> letter to this patch series https://lkml.org/lkml/2013/3/9/73

Well well... Either you think this must be done and isn't questionable,
and you shouldn't bother Cc'ing a dozen driver maintainers and waiting
for them to ack. Or you think this is something for discussion and you
should consistently Cc all interested parties on the whole patch
series. If you send me one random patch in the middle of a series and
expect me to go fish for the missing parts so that I can make sense of
it all and make sane and useful comments, well this isn't going to
happen, sorry.

Now with the above pointers, I can actually make useful comments, and I
will.

-- 
Jean Delvare