Re: [PATCH net-next v9 3/3] net: dsa: yt921x: Add support for Motorcomm YT921x

[Andrew Lunn]

> I'm not sure that a "driver lock" is something that drivers need.
> In this case it creates a lot of red tape. Function yt921x_dsa_X() takes
> the driver lock and calls function yt921x_X() which does the work.
> IMO that's part of what gives "vendor crap" drivers their name, when
> there's no reason behind it.

As you said, some methods are protected by RTNL. But not all. And it
is hard to know which are not. A driver lock is KISS, and easy to get
right, easy to see is right, and easy to prove is right. Locking can
be hard, so KISS is good.

	Andrew