Messages in this thread | | | Subject | Re: [PATCH v7 1/4] spinlock: A new lockref structure for lockless update of refcount | From | Benjamin Herrenschmidt <> | Date | Fri, 30 Aug 2013 10:26:48 +1000 |
| |
On Thu, 2013-08-29 at 16:42 -0700, Linus Torvalds wrote:
> For architecture people (ie Ben, if you want to try this on ppc64), > the thing that it needs from an architecture: > > - the raw_spinlock_t and the "unsigned long" needs to fit in a u64.
I assume you mean unsigned int ? :-)
.../...
> - the architecture needs to implement a simple > "arch_spin_value_unlocked()" macro, which takes a raw_spinlock_t value > and says whether it is unlocked or not.
What's wrong with the existing arch_spin_is_locked() ?
BTW. Do you have your test case at hand ?
Cheers, Ben.
| |