Re: [PATCH v3 071/114] lockdep: add lockdep_assert_not_held

[Peter Zijlstra <peterz@xxxxxxxxxxxxx>]

On Tue, Jul 01, 2014 at 06:41:11AM -0400, Jeff Layton wrote:
> Right -- in the case of something like atomic_dec_and_lock, we only
> take the spinlock if we think the count might go to zero. So, we might
> miss catching some places that could deadlock if the refcounts don't go
> to zero in the testing we're doing.
> 
> might_lock may be what we need, but I don't see any callers of it, 

might_fault()
  might_lock_read(&current->mm->mmap_sem)

I'm not sure we have others; but this is the one I remember providing
this for.

> and
> at a quick glance it doesn't appear to be disabled if debug_locks is
> false.

The whole of lockdep gets killed, and that would include the
lock_acquire/lock_release used to implement these.

Attachment: pgpYx2m0KXatY.pgp
Description: PGP signature