Palacios Public Git Repository

To checkout Palacios execute

  git clone http://v3vee.org/palacios/palacios.web/palacios.git
This will give you the master branch. You probably want the devel branch or one of the release branches. To switch to the devel branch, simply execute
  cd palacios
  git checkout --track -b devel origin/devel
The other branches are similar.


Merge branch 'devel'
[palacios.git] / kitten / include / arch-x86_64 / spinlock.h
diff --git a/kitten/include/arch-x86_64/spinlock.h b/kitten/include/arch-x86_64/spinlock.h
new file mode 100644 (file)
index 0000000..b3f1256
--- /dev/null
@@ -0,0 +1,133 @@
+#ifndef _X86_64_SPINLOCK_H
+#define _X86_64_SPINLOCK_H
+
+#include <arch/atomic.h>
+#include <arch/rwlock.h>
+#include <arch/page.h>
+
+/*
+ * Your basic SMP spinlocks, allowing only a single CPU anywhere
+ *
+ * Simple spin lock operations.  There are two variants, one clears IRQ's
+ * on the local processor, one does not.
+ *
+ * We make no fairness assumptions. They have a cost.
+ *
+ * (the type definitions are in arch/spinlock_types.h)
+ */
+
+#define __raw_spin_is_locked(x) \
+               (*(volatile signed int *)(&(x)->slock) <= 0)
+
+#define __raw_spin_lock_string \
+       "\n1:\t" \
+       "lock ; decl %0\n\t" \
+       "js 2f\n" \
+       LOCK_SECTION_START("") \
+       "2:\t" \
+       "rep;nop\n\t" \
+       "cmpl $0,%0\n\t" \
+       "jle 2b\n\t" \
+       "jmp 1b\n" \
+       LOCK_SECTION_END
+
+#define __raw_spin_unlock_string \
+       "movl $1,%0" \
+               :"=m" (lock->slock) : : "memory"
+
+static inline void __raw_spin_lock(raw_spinlock_t *lock)
+{
+       __asm__ __volatile__(
+               __raw_spin_lock_string
+               :"=m" (lock->slock) : : "memory");
+}
+
+#define __raw_spin_lock_flags(lock, flags) __raw_spin_lock(lock)
+
+static inline int __raw_spin_trylock(raw_spinlock_t *lock)
+{
+       int oldval;
+
+       __asm__ __volatile__(
+               "xchgl %0,%1"
+               :"=q" (oldval), "=m" (lock->slock)
+               :"0" (0) : "memory");
+
+       return oldval > 0;
+}
+
+static inline void __raw_spin_unlock(raw_spinlock_t *lock)
+{
+       __asm__ __volatile__(
+               __raw_spin_unlock_string
+       );
+}
+
+#define __raw_spin_unlock_wait(lock) \
+       do { while (__raw_spin_is_locked(lock)) cpu_relax(); } while (0)
+
+/*
+ * Read-write spinlocks, allowing multiple readers
+ * but only one writer.
+ *
+ * NOTE! it is quite common to have readers in interrupts
+ * but no interrupt writers. For those circumstances we
+ * can "mix" irq-safe locks - any writer needs to get a
+ * irq-safe write-lock, but readers can get non-irqsafe
+ * read-locks.
+ *
+ * On x86, we implement read-write locks as a 32-bit counter
+ * with the high bit (sign) being the "contended" bit.
+ *
+ * The inline assembly is non-obvious. Think about it.
+ *
+ * Changed to use the same technique as rw semaphores.  See
+ * semaphore.h for details.  -ben
+ *
+ * the helpers are in arch/i386/kernel/semaphore.c
+ */
+
+#define __raw_read_can_lock(x)         ((int)(x)->lock > 0)
+#define __raw_write_can_lock(x)                ((x)->lock == RW_LOCK_BIAS)
+
+static inline void __raw_read_lock(raw_rwlock_t *rw)
+{
+       __build_read_lock(rw, "__read_lock_failed");
+}
+
+static inline void __raw_write_lock(raw_rwlock_t *rw)
+{
+       __build_write_lock(rw, "__write_lock_failed");
+}
+
+static inline int __raw_read_trylock(raw_rwlock_t *lock)
+{
+       atomic_t *count = (atomic_t *)lock;
+       atomic_dec(count);
+       if (atomic_read(count) >= 0)
+               return 1;
+       atomic_inc(count);
+       return 0;
+}
+
+static inline int __raw_write_trylock(raw_rwlock_t *lock)
+{
+       atomic_t *count = (atomic_t *)lock;
+       if (atomic_sub_and_test(RW_LOCK_BIAS, count))
+               return 1;
+       atomic_add(RW_LOCK_BIAS, count);
+       return 0;
+}
+
+static inline void __raw_read_unlock(raw_rwlock_t *rw)
+{
+       asm volatile("lock ; incl %0" :"=m" (rw->lock) : : "memory");
+}
+
+static inline void __raw_write_unlock(raw_rwlock_t *rw)
+{
+       asm volatile("lock ; addl $" RW_LOCK_BIAS_STR ",%0"
+                               : "=m" (rw->lock) : : "memory");
+}
+
+#endif /* _X86_64_SPINLOCK_H */