From: Zwane Mwaikambo <zwane@linuxpower.ca>

Following up on Keith's code, I adapted the i386 code to allow enabling
interrupts during contested locks depending on previous interrupt
enable status. Obviously there will be a text increase (only for non
CONFIG_SPINLINE case), although it doesn't seem so bad, there will be an
increased exit latency when we attempt a lock acquisition after spinning
due to the extra instructions. How much this will affect performance I'm
not sure yet as I haven't had time to micro bench.

   text    data     bss     dec     hex filename
2628024  921731       0 3549755  362a3b vmlinux-after
2621369  921731       0 3543100  36103c vmlinux-before
2618313  919222       0 3537535  35fa7f vmlinux-spinline

The code has been stress tested on a 16x NUMAQ (courtesy OSDL).


---

 25-akpm/include/asm-i386/spinlock.h |   47 +++++++++++++++++++++++++++++++++++-
 1 files changed, 46 insertions(+), 1 deletion(-)

diff -puN include/asm-i386/spinlock.h~allow-i386-to-reenable-interrupts-on-lock-contention include/asm-i386/spinlock.h
--- 25/include/asm-i386/spinlock.h~allow-i386-to-reenable-interrupts-on-lock-contention	2004-05-01 15:47:19.211477648 -0700
+++ 25-akpm/include/asm-i386/spinlock.h	2004-05-01 15:47:19.215477040 -0700
@@ -42,7 +42,6 @@ typedef struct {
 
 #define spin_is_locked(x)	(*(volatile signed char *)(&(x)->lock) <= 0)
 #define spin_unlock_wait(x)	do { barrier(); } while(spin_is_locked(x))
-#define _raw_spin_lock_flags(lock, flags) _raw_spin_lock(lock)
 
 #ifdef CONFIG_SPINLINE
 
@@ -58,6 +57,21 @@ typedef struct {
 		"jmp 1b\n" \
 		"3:\t"
 
+	#define spin_lock_string_flags \
+		"\n1:\t" \
+		"lock ; decb %0\n\t" \
+		"jns 3f\n" \
+		"testl $0x200, %1\n\t" \
+		"jz 2f\n\t" \
+		"sti\n\t" \
+		"2:\t" \
+		"rep;nop\n\t" \
+		"cmpb $0, %0\n\t" \
+		"jle 2b\n\t" \
+		"cli\n\t" \
+		"jmp 1b\n" \
+		"3:\t"
+
 #else /* !CONFIG_SPINLINE */
 
 	#define spin_lock_string \
@@ -72,6 +86,23 @@ typedef struct {
 		"jmp 1b\n" \
 		LOCK_SECTION_END
 
+	#define spin_lock_string_flags \
+		"\n1:\t" \
+		"lock ; decb %0\n\t" \
+		"js 2f\n\t" \
+		LOCK_SECTION_START("") \
+		"2:\t" \
+		"testl $0x200, %1\n\t" \
+		"jz 3f\n\t" \
+		"sti\n\t" \
+		"3:\t" \
+		"rep;nop\n\t" \
+		"cmpb $0, %0\n\t" \
+		"jle 3b\n\t" \
+		"cli\n\t" \
+		"jmp 1b\n" \
+		LOCK_SECTION_END
+
 #endif /* CONFIG_SPINLINE */
 /*
  * This works. Despite all the confusion.
@@ -143,6 +174,20 @@ here:
 		:"=m" (lock->lock) : : "memory");
 }
 
+static inline void _raw_spin_lock_flags (spinlock_t *lock, unsigned long flags)
+{
+#ifdef CONFIG_DEBUG_SPINLOCK
+	__label__ here;
+here:
+	if (unlikely(lock->magic != SPINLOCK_MAGIC)) {
+		printk("eip: %p\n", &&here);
+		BUG();
+	}
+#endif
+	__asm__ __volatile__(
+		spin_lock_string_flags
+		:"=m" (lock->lock) : "r" (flags) : "memory");
+}
 
 /*
  * Read-write spinlocks, allowing multiple readers

_