On Fri, Nov 14, 2014 at 7:09 PM, Suman Anna [off-list ref] wrote:

quoted

It seems to me that hwspin_lock_request_specific failures should be
used by clients to defer their probing. Why wouldn't such a simple
solution work?

Because the API always returns NULL on failures and there is no way for
the clients to figure out if the lock id passed is invalid or the bank
containing the lock is not registered.

It seems to me this may always be the case - lock ids may be wrong and
there's no way to fully validate them.

Let's start with the simpler approach where
hwspin_lock_request_specific failures are used by clients to defer
their probing.

If a real use case will require changes we can always do that later,
though it seems to me the only gain by changing this API is to catch a
small subset of fatal DT mistakes which will anyway be caught very
early in the development.

Thanks,
Ohad.