Re: [patch 1/3] spinlock fix #1, *_can_lock() primitives

[Ingo Molnar <mingo@elte.hu>]

* Linus Torvalds <torvalds@osdl.org> wrote:

> I can do ppc64 myself, can others fix the other architectures (Ingo,
> shouldn't the UP case have the read/write_can_lock() cases too? And
> wouldn't you agree that it makes more sense to have the rwlock test
> variants in asm/rwlock.h?):

this one adds it to x64. (untested at the moment) This patch assumes
that we are nuking rwlock_is_locked and that there is at least a
s/rwlock_is_locked/!write_can_lock/ done to kernel/exit.c.

	Ingo

Signed-off-by: Ingo Molnar <mingo@elte.hu>

--- linux/include/asm-x86_64/spinlock.h.orig
+++ linux/include/asm-x86_64/spinlock.h
@@ -161,7 +161,23 @@ typedef struct {
 
 #define rwlock_init(x)	do { *(x) = RW_LOCK_UNLOCKED; } while(0)
 
-#define rwlock_is_locked(x) ((x)->lock != RW_LOCK_BIAS)
+/**
+ * read_can_lock - would read_trylock() succeed?
+ * @lock: the rwlock in question.
+ */
+static inline int read_can_lock(rwlock_t *rw)
+{
+	return rw->lock > 0;
+}
+
+/**
+ * write_can_lock - would write_trylock() succeed?
+ * @lock: the rwlock in question.
+ */
+static inline int write_can_lock(rwlock_t *rw)
+{
+	return rw->lock == RW_LOCK_BIAS;
+}
 
 /*
  * On x86, we implement read-write locks as a 32-bit counter