Re: Converting dev->mutex into dev->spinlock ?

[Greg Kroah-Hartman]

On Sat, Feb 04, 2023 at 10:32:11PM +0900, Tetsuo Handa wrote:
> Hello.
> 
> There is a long-standing deadlock problem in driver core code caused by
> "struct device"->mutex being marked as "do not apply lockdep checks".

The marking of a lock does not cause a deadlock problem, so what do you
mean exactly by this?  Where is the actual deadlock?

> We can make this deadlock visible by applying [1], and we can confirm that
> there is a deadlock problem that I think needs to be addressed in core code [2].

Any reason why you didn't cc: us on these patches?

> Also, since driver developers are taking it for granted that driver callback
> functions can behave as if dev->mutex is not held (because possibility of deadlock
> was never reported), it would solve many deadlocks in driver code if you can update
> driver core code to avoid calling driver callback functions with dev->mutex held
> (by e.g. replacing dev->mutex with dev->spinlock and dev->atomic_flags).
> But I'm not familiar enough to propose such change...

A driver developer should never be messing with the mutex of a device,
that's not for them to touch, that's the driver core's lock to touch,
right?

So I don't understand the real problem here.  What subsystem is having
issues and what issues are they exactly?

And using a spinlock shouldn't change any locking deadlocks that I can
tell, so I don't understand the proposal, sorry.

thanks,

greg k-h