Is spin_is_locked() safe to use with BUG_ON()/WARN_ON()?

[David Howells <dhowells@redhat.com>]

We are using spin_is_locked() in a few places to give a warning or an oops if
either a spinlock is not held or if it is held.  I'm not sure all of these are
safe.

Take uas_try_complete() in drivers/usb/storage/uas.c which does:

	WARN_ON(!spin_is_locked(&devinfo->lock));

or fscache_start_operations() which does:

	ASSERT(spin_is_locked(&object->lock));

These will unconditionally fail under sometimes because under certain
conditions spin_is_locked() is hardwired to 0 (ie. not locked) when actually
we're in a place where the spinlock _should_ be locked, and we should get a
non-zero return.


Would it be reasonable to add a spin_is_not_locked() function for use when we
expect it not to be locked and then use spin_is_locked() only when we expect it
to be locked?

Thanks to Milosz Tanski for spotting this one.

David