arm64: klib: Optimised atomic bitops

This patch implements the AArch64-specific atomic bitops functions using
exclusive memory accesses to avoid locking.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>

arch/arm64/include/asm/bitops.h

@@ -32,6 +32,16 @@
 #error only <linux/bitops.h> can be included directly
 #endif
 
+/*
+ * Little endian assembly atomic bitops.
+ */
+extern void set_bit(int nr, volatile unsigned long *p);
+extern void clear_bit(int nr, volatile unsigned long *p);
+extern void change_bit(int nr, volatile unsigned long *p);
+extern int test_and_set_bit(int nr, volatile unsigned long *p);
+extern int test_and_clear_bit(int nr, volatile unsigned long *p);
+extern int test_and_change_bit(int nr, volatile unsigned long *p);
+
 #include <asm-generic/bitops/builtin-__ffs.h>
 #include <asm-generic/bitops/builtin-ffs.h>
 #include <asm-generic/bitops/builtin-__fls.h>
@@ -45,9 +55,13 @@
 #include <asm-generic/bitops/hweight.h>
 #include <asm-generic/bitops/lock.h>
 
-#include <asm-generic/bitops/atomic.h>
 #include <asm-generic/bitops/non-atomic.h>
 #include <asm-generic/bitops/le.h>
-#include <asm-generic/bitops/ext2-atomic.h>
+
+/*
+ * Ext2 is defined to use little-endian byte ordering.
+ */
+#define ext2_set_bit_atomic(lock, nr, p)	test_and_set_bit_le(nr, p)
+#define ext2_clear_bit_atomic(lock, nr, p)	test_and_clear_bit_le(nr, p)
 
 #endif /* __ASM_BITOPS_H */

arch/arm64/kernel/arm64ksyms.c

@@ -39,11 +39,6 @@ EXPORT_SYMBOL(__copy_from_user);
 EXPORT_SYMBOL(__copy_to_user);
 EXPORT_SYMBOL(__clear_user);
 
-	/* bitops */
-#ifdef CONFIG_SMP
-EXPORT_SYMBOL(__atomic_hash);
-#endif
-
 	/* physical memory */
 EXPORT_SYMBOL(memstart_addr);
 
@@ -54,3 +49,11 @@ EXPORT_SYMBOL(memset);
 EXPORT_SYMBOL(memcpy);
 EXPORT_SYMBOL(memmove);
 EXPORT_SYMBOL(memchr);
+
+	/* atomic bitops */
+EXPORT_SYMBOL(set_bit);
+EXPORT_SYMBOL(test_and_set_bit);
+EXPORT_SYMBOL(clear_bit);
+EXPORT_SYMBOL(test_and_clear_bit);
+EXPORT_SYMBOL(change_bit);
+EXPORT_SYMBOL(test_and_change_bit);

arch/arm64/lib/bitops.c → arch/arm64/lib/bitops.S

 /*
- * Copyright (C) 2012 ARM Limited
+ * Based on arch/arm/lib/bitops.h
+ *
+ * Copyright (C) 2013 ARM Ltd.
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 2 as
@@ -14,12 +16,55 @@
  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
-#include <linux/kernel.h>
-#include <linux/spinlock.h>
-#include <linux/atomic.h>
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+/*
+ * x0: bits 5:0  bit offset
+ *     bits 63:6 word offset
+ * x1: address
+ */
+	.macro	bitop, name, instr
+ENTRY(	\name	)
+	and	x3, x0, #63		// Get bit offset
+	eor	x0, x0, x3		// Clear low bits
+	mov	x2, #1
+	add	x1, x1, x0, lsr #3	// Get word offset
+	lsl	x3, x2, x3		// Create mask
+1:	ldxr	x2, [x1]
+	\instr	x2, x2, x3
+	stxr	w0, x2, [x1]
+	cbnz	w0, 1b
+	ret
+ENDPROC(\name	)
+	.endm
+
+	.macro	testop, name, instr
+ENTRY(	\name	)
+	and	x3, x0, #63		// Get bit offset
+	eor	x0, x0, x3		// Clear low bits
+	mov	x2, #1
+	add	x1, x1, x0, lsr #3	// Get word offset
+	lsl	x4, x2, x3		// Create mask
+	smp_dmb	ish
+1:	ldxr	x2, [x1]
+	lsr	x0, x2, x3		// Save old value of bit
+	\instr	x2, x2, x4		// toggle bit
+	stxr	w2, x2, [x1]
+	cbnz	w2, 1b
+	smp_dmb	ish
+	and	x0, x0, #1
+3:	ret
+ENDPROC(\name	)
+	.endm
+
+/*
+ * Atomic bit operations.
+ */
+	bitop	change_bit, eor
+	bitop	clear_bit, bic
+	bitop	set_bit, orr
 
-#ifdef CONFIG_SMP
-arch_spinlock_t __atomic_hash[ATOMIC_HASH_SIZE] __lock_aligned = {
-       [0 ... (ATOMIC_HASH_SIZE-1)]  = __ARCH_SPIN_LOCK_UNLOCKED
-};
-#endif
+	testop	test_and_change_bit, eor
+	testop	test_and_clear_bit, bic
+	testop	test_and_set_bit, orr