sparc support code for librthread (_atomic_lock yanked from existing

libpthread code).

1 files changed, 41 insertions, 0 deletions

diff --git a/lib/librthread/arch/sparc/_atomic_lock.c b/lib/librthread/arch/sparc/_atomic_lock.c
new file mode 100644
index 00000000000..c791314c8d1
--- /dev/null
+++ b/lib/librthread/arch/sparc/_atomic_lock.c
@@ -0,0 +1,41 @@
+/*	$OpenBSD: _atomic_lock.c,v 1.1 2005/12/25 16:00:02 miod Exp $	*/
+/*
+ * Atomic lock for sparc
+ */
+ 
+#include <machine/spinlock.h>
+
+int
+_atomic_lock(volatile _spinlock_lock_t * lock)
+{
+	_spinlock_lock_t old;
+
+	/*
+	 *  "  ldstub  [address], reg_rd
+	 * 
+	 *  The atomic load-store instructions copy a byte from memory
+	 *  into r[rd]m then rewrite the addressed byte in memory to all
+	 *  ones [_SPINLOCK_LOCKED]. The operation is performed
+	 *  atomically, that is, without allowing intervening interrupts
+	 *  or deferred traps. In a multiprocessor system, two or more
+	 *  processors executing atomic load-store unsigned byte [...]
+	 *  addressing the same byte [...] simultaneously are guaranteed
+	 *  to execute them in an undefined, but serial order."
+	 *    - p101, The SPARC Architecture Manual (version 8) Prentice-Hall
+	 *
+	 * "LDSTUB loads a byte value from memory to a register and writes
+	 *  the value FF_16 into the addressed byte atomically. LDSTUB
+	 *  is the classic test-and-set instruction. Like SWAP, it has
+	 *  a consensus number of two and so cannot resolve more than
+	 *  two contending processes in a wait-free fashion."
+	 *    - p129, The SPARC Architecture Manual (version 9) Prentice-Hall
+	 *  (See also section J.6 (spinlocks))
+	 *
+	 * (No change to the condition codes are documented.)
+	 */
+	__asm__("ldstub %0,%1"
+		: "=m" (*lock), "=r" (old)
+		: "0" (*lock));
+
+	return (old == _SPINLOCK_LOCKED);
+}