Hello Thorsten,

On Fri, Feb 14, 2020 at 12:02:38PM +0100, Thorsten Scherer wrote:

AFAICT this is all good.

Unfortunately i don't have any idea on how to test out the difference
this patch makes on a real SIOX.

When I started looking into the problem I expected that the lock
debugging would catch these problems. But either I did something wrong
or there is no mechanism that catches

 - a spinlock is taken when there is already a raw spinlock taken
 - spin_lock_irq is used with irqs already off

And I wonder if there are reasons I don't see that make these two tests
a bad idea.

Any hints? Is it necessary at all?

Apart from "normal" testing that SIOX still works I have no good
suggestion. Having said that I would be surprised if my patch breaked
something. (But it wouldn't be the first time such a surprise happens
:-)

Best regards
Uwe
-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | https://www.pengutronix.de/ |