Blackfin arch: SMP supporting patchset: Blackfin header files and machine common code

Blackfin dual core BF561 processor can support SMP like features.
https://docs.blackfin.uclinux.org/doku.php?id=linux-kernel:smp-like

In this patch, we provide SMP extend to Blackfin header files
and machine common code

Signed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

18 files changed, 829 insertions, 121 deletions

diff --git a/arch/blackfin/include/asm/atomic.h b/arch/blackfin/include/asm/atomic.h
index 25776c19064..d76275e5638 100644
--- a/arch/blackfin/include/asm/atomic.h
+++ b/arch/blackfin/include/asm/atomic.h
@@ -15,11 +15,80 @@
  */
 
 #define ATOMIC_INIT(i)	{ (i) }
-
-#define atomic_read(v)		((v)->counter)
 #define atomic_set(v, i)	(((v)->counter) = i)
 
-static __inline__ void atomic_add(int i, atomic_t * v)
+#ifdef CONFIG_SMP
+
+#define atomic_read(v)	__raw_uncached_fetch_asm(&(v)->counter)
+
+asmlinkage int __raw_uncached_fetch_asm(const volatile int *ptr);
+
+asmlinkage int __raw_atomic_update_asm(volatile int *ptr, int value);
+
+asmlinkage int __raw_atomic_clear_asm(volatile int *ptr, int value);
+
+asmlinkage int __raw_atomic_set_asm(volatile int *ptr, int value);
+
+asmlinkage int __raw_atomic_xor_asm(volatile int *ptr, int value);
+
+asmlinkage int __raw_atomic_test_asm(const volatile int *ptr, int value);
+
+static inline void atomic_add(int i, atomic_t *v)
+{
+	__raw_atomic_update_asm(&v->counter, i);
+}
+
+static inline void atomic_sub(int i, atomic_t *v)
+{
+	__raw_atomic_update_asm(&v->counter, -i);
+}
+
+static inline int atomic_add_return(int i, atomic_t *v)
+{
+	return __raw_atomic_update_asm(&v->counter, i);
+}
+
+static inline int atomic_sub_return(int i, atomic_t *v)
+{
+	return __raw_atomic_update_asm(&v->counter, -i);
+}
+
+static inline void atomic_inc(volatile atomic_t *v)
+{
+	__raw_atomic_update_asm(&v->counter, 1);
+}
+
+static inline void atomic_dec(volatile atomic_t *v)
+{
+	__raw_atomic_update_asm(&v->counter, -1);
+}
+
+static inline void atomic_clear_mask(int mask, atomic_t *v)
+{
+	__raw_atomic_clear_asm(&v->counter, mask);
+}
+
+static inline void atomic_set_mask(int mask, atomic_t *v)
+{
+	__raw_atomic_set_asm(&v->counter, mask);
+}
+
+static inline int atomic_test_mask(int mask, atomic_t *v)
+{
+	return __raw_atomic_test_asm(&v->counter, mask);
+}
+
+/* Atomic operations are already serializing */
+#define smp_mb__before_atomic_dec()    barrier()
+#define smp_mb__after_atomic_dec() barrier()
+#define smp_mb__before_atomic_inc()    barrier()
+#define smp_mb__after_atomic_inc() barrier()
+
+#else /* !CONFIG_SMP */
+
+#define atomic_read(v)	((v)->counter)
+
+static inline void atomic_add(int i, atomic_t *v)
 {
 	long flags;
 
@@ -28,7 +97,7 @@ static __inline__ void atomic_add(int i, atomic_t * v)
 	local_irq_restore(flags);
 }
 
-static __inline__ void atomic_sub(int i, atomic_t * v)
+static inline void atomic_sub(int i, atomic_t *v)
 {
 	long flags;
 
@@ -38,7 +107,7 @@ static __inline__ void atomic_sub(int i, atomic_t * v)
 
 }
 
-static inline int atomic_add_return(int i, atomic_t * v)
+static inline int atomic_add_return(int i, atomic_t *v)
 {
 	int __temp = 0;
 	long flags;
@@ -52,8 +121,7 @@ static inline int atomic_add_return(int i, atomic_t * v)
 	return __temp;
 }
 
-#define atomic_add_negative(a, v)	(atomic_add_return((a), (v)) < 0)
-static inline int atomic_sub_return(int i, atomic_t * v)
+static inline int atomic_sub_return(int i, atomic_t *v)
 {
 	int __temp = 0;
 	long flags;
@@ -66,7 +134,7 @@ static inline int atomic_sub_return(int i, atomic_t * v)
 	return __temp;
 }
 
-static __inline__ void atomic_inc(volatile atomic_t * v)
+static inline void atomic_inc(volatile atomic_t *v)
 {
 	long flags;
 
@@ -75,20 +143,7 @@ static __inline__ void atomic_inc(volatile atomic_t * v)
 	local_irq_restore(flags);
 }
 
-#define atomic_cmpxchg(v, o, n) ((int)cmpxchg(&((v)->counter), (o), (n)))
-#define atomic_xchg(v, new) (xchg(&((v)->counter), new))
-
-#define atomic_add_unless(v, a, u)				\
-({								\
-	int c, old;						\
-	c = atomic_read(v);					\
-	while (c != (u) && (old = atomic_cmpxchg((v), c, c + (a))) != c) \
-		c = old;					\
-	c != (u);						\
-})
-#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0)
-
-static __inline__ void atomic_dec(volatile atomic_t * v)
+static inline void atomic_dec(volatile atomic_t *v)
 {
 	long flags;
 
@@ -97,7 +152,7 @@ static __inline__ void atomic_dec(volatile atomic_t * v)
 	local_irq_restore(flags);
 }
 
-static __inline__ void atomic_clear_mask(unsigned int mask, atomic_t * v)
+static inline void atomic_clear_mask(unsigned int mask, atomic_t *v)
 {
 	long flags;
 
@@ -106,7 +161,7 @@ static __inline__ void atomic_clear_mask(unsigned int mask, atomic_t * v)
 	local_irq_restore(flags);
 }
 
-static __inline__ void atomic_set_mask(unsigned int mask, atomic_t * v)
+static inline void atomic_set_mask(unsigned int mask, atomic_t *v)
 {
 	long flags;
 
@@ -121,9 +176,25 @@ static __inline__ void atomic_set_mask(unsigned int mask, atomic_t * v)
 #define smp_mb__before_atomic_inc()    barrier()
 #define smp_mb__after_atomic_inc() barrier()
 
+#endif /* !CONFIG_SMP */
+
+#define atomic_add_negative(a, v)	(atomic_add_return((a), (v)) < 0)
 #define atomic_dec_return(v) atomic_sub_return(1,(v))
 #define atomic_inc_return(v) atomic_add_return(1,(v))
 
+#define atomic_cmpxchg(v, o, n) ((int)cmpxchg(&((v)->counter), (o), (n)))
+#define atomic_xchg(v, new) (xchg(&((v)->counter), new))
+
+#define atomic_add_unless(v, a, u)				\
+({								\
+	int c, old;						\
+	c = atomic_read(v);					\
+	while (c != (u) && (old = atomic_cmpxchg((v), c, c + (a))) != c) \
+		c = old;					\
+	c != (u);						\
+})
+#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0)
+
 /*
  * atomic_inc_and_test - increment and test
  * @v: pointer of type atomic_t
diff --git a/arch/blackfin/include/asm/bfin-global.h b/arch/blackfin/include/asm/bfin-global.h
index 77295666c34..1dd08058bc9 100644
--- a/arch/blackfin/include/asm/bfin-global.h
+++ b/arch/blackfin/include/asm/bfin-global.h
@@ -47,6 +47,9 @@
 # define DMA_UNCACHED_REGION (0)
 #endif
 
+extern void bfin_setup_caches(unsigned int cpu);
+extern void bfin_setup_cpudata(unsigned int cpu);
+
 extern unsigned long get_cclk(void);
 extern unsigned long get_sclk(void);
 extern unsigned long sclk_to_usecs(unsigned long sclk);
@@ -58,8 +61,6 @@ extern void dump_bfin_trace_buffer(void);
 
 /* init functions only */
 extern int init_arch_irq(void);
-extern void bfin_icache_init(void);
-extern void bfin_dcache_init(void);
 extern void init_exception_vectors(void);
 extern void program_IAR(void);
 
diff --git a/arch/blackfin/include/asm/bitops.h b/arch/blackfin/include/asm/bitops.h
index c428e4106f8..9964e17232e 100644
--- a/arch/blackfin/include/asm/bitops.h
+++ b/arch/blackfin/include/asm/bitops.h
@@ -7,7 +7,6 @@
 
 #include <linux/compiler.h>
 #include <asm/byteorder.h>	/* swab32 */
-#include <asm/system.h>		/* save_flags */
 
 #ifdef __KERNEL__
 
@@ -20,36 +19,71 @@
 #include <asm-generic/bitops/sched.h>
 #include <asm-generic/bitops/ffz.h>
 
-static __inline__ void set_bit(int nr, volatile unsigned long *addr)
+#ifdef CONFIG_SMP
+
+#include <linux/linkage.h>
+
+asmlinkage int __raw_bit_set_asm(volatile unsigned long *addr, int nr);
+
+asmlinkage int __raw_bit_clear_asm(volatile unsigned long *addr, int nr);
+
+asmlinkage int __raw_bit_toggle_asm(volatile unsigned long *addr, int nr);
+
+asmlinkage int __raw_bit_test_set_asm(volatile unsigned long *addr, int nr);
+
+asmlinkage int __raw_bit_test_clear_asm(volatile unsigned long *addr, int nr);
+
+asmlinkage int __raw_bit_test_toggle_asm(volatile unsigned long *addr, int nr);
+
+asmlinkage int __raw_bit_test_asm(const volatile unsigned long *addr, int nr);
+
+static inline void set_bit(int nr, volatile unsigned long *addr)
 {
-	int *a = (int *)addr;
-	int mask;
-	unsigned long flags;
+	volatile unsigned long *a = addr + (nr >> 5);
+	__raw_bit_set_asm(a, nr & 0x1f);
+}
 
-	a += nr >> 5;
-	mask = 1 << (nr & 0x1f);
-	local_irq_save(flags);
-	*a |= mask;
-	local_irq_restore(flags);
+static inline void clear_bit(int nr, volatile unsigned long *addr)
+{
+	volatile unsigned long *a = addr + (nr >> 5);
+	__raw_bit_clear_asm(a, nr & 0x1f);
 }
 
-static __inline__ void __set_bit(int nr, volatile unsigned long *addr)
+static inline void change_bit(int nr, volatile unsigned long *addr)
 {
-	int *a = (int *)addr;
-	int mask;
+	volatile unsigned long *a = addr + (nr >> 5);
+	__raw_bit_toggle_asm(a, nr & 0x1f);
+}
 
-	a += nr >> 5;
-	mask = 1 << (nr & 0x1f);
-	*a |= mask;
+static inline int test_bit(int nr, const volatile unsigned long *addr)
+{
+	volatile const unsigned long *a = addr + (nr >> 5);
+	return __raw_bit_test_asm(a, nr & 0x1f) != 0;
 }
 
-/*
- * clear_bit() doesn't provide any barrier for the compiler.
- */
-#define smp_mb__before_clear_bit()	barrier()
-#define smp_mb__after_clear_bit()	barrier()
+static inline int test_and_set_bit(int nr, volatile unsigned long *addr)
+{
+	volatile unsigned long *a = addr + (nr >> 5);
+	return __raw_bit_test_set_asm(a, nr & 0x1f);
+}
 
-static __inline__ void clear_bit(int nr, volatile unsigned long *addr)
+static inline int test_and_clear_bit(int nr, volatile unsigned long *addr)
+{
+	volatile unsigned long *a = addr + (nr >> 5);
+	return __raw_bit_test_clear_asm(a, nr & 0x1f);
+}
+
+static inline int test_and_change_bit(int nr, volatile unsigned long *addr)
+{
+	volatile unsigned long *a = addr + (nr >> 5);
+	return __raw_bit_test_toggle_asm(a, nr & 0x1f);
+}
+
+#else /* !CONFIG_SMP */
+
+#include <asm/system.h>		/* save_flags */
+
+static inline void set_bit(int nr, volatile unsigned long *addr)
 {
 	int *a = (int *)addr;
 	int mask;
@@ -57,21 +91,23 @@ static __inline__ void clear_bit(int nr, volatile unsigned long *addr)
 	a += nr >> 5;
 	mask = 1 << (nr & 0x1f);
 	local_irq_save(flags);
-	*a &= ~mask;
+	*a |= mask;
 	local_irq_restore(flags);
 }
 
-static __inline__ void __clear_bit(int nr, volatile unsigned long *addr)
+static inline void clear_bit(int nr, volatile unsigned long *addr)
 {
 	int *a = (int *)addr;
 	int mask;
-
+	unsigned long flags;
 	a += nr >> 5;
 	mask = 1 << (nr & 0x1f);
+	local_irq_save(flags);
 	*a &= ~mask;
+	local_irq_restore(flags);
 }
 
-static __inline__ void change_bit(int nr, volatile unsigned long *addr)
+static inline void change_bit(int nr, volatile unsigned long *addr)
 {
 	int mask, flags;
 	unsigned long *ADDR = (unsigned long *)addr;
@@ -83,17 +119,7 @@ static __inline__ void change_bit(int nr, volatile unsigned long *addr)
 	local_irq_restore(flags);
 }
 
-static __inline__ void __change_bit(int nr, volatile unsigned long *addr)
-{
-	int mask;
-	unsigned long *ADDR = (unsigned long *)addr;
-
-	ADDR += nr >> 5;
-	mask = 1 << (nr & 31);
-	*ADDR ^= mask;
-}
-
-static __inline__ int test_and_set_bit(int nr, void *addr)
+static inline int test_and_set_bit(int nr, volatile unsigned long *addr)
 {
 	int mask, retval;
 	volatile unsigned int *a = (volatile unsigned int *)addr;
@@ -109,19 +135,23 @@ static __inline__ int test_and_set_bit(int nr, void *addr)
 	return retval;
 }
 
-static __inline__ int __test_and_set_bit(int nr, volatile unsigned long *addr)
+static inline int test_and_clear_bit(int nr, volatile unsigned long *addr)
 {
 	int mask, retval;
 	volatile unsigned int *a = (volatile unsigned int *)addr;
+	unsigned long flags;
 
 	a += nr >> 5;
 	mask = 1 << (nr & 0x1f);
+	local_irq_save(flags);
 	retval = (mask & *a) != 0;
-	*a |= mask;
+	*a &= ~mask;
+	local_irq_restore(flags);
+
 	return retval;
 }
 
-static __inline__ int test_and_clear_bit(int nr, volatile unsigned long *addr)
+static inline int test_and_change_bit(int nr, volatile unsigned long *addr)
 {
 	int mask, retval;
 	volatile unsigned int *a = (volatile unsigned int *)addr;
@@ -131,13 +161,50 @@ static __inline__ int test_and_clear_bit(int nr, volatile unsigned long *addr)
 	mask = 1 << (nr & 0x1f);
 	local_irq_save(flags);
 	retval = (mask & *a) != 0;
-	*a &= ~mask;
+	*a ^= mask;
 	local_irq_restore(flags);
-
 	return retval;
 }
 
-static __inline__ int __test_and_clear_bit(int nr, volatile unsigned long *addr)
+#endif /* CONFIG_SMP */
+
+/*
+ * clear_bit() doesn't provide any barrier for the compiler.
+ */
+#define smp_mb__before_clear_bit()	barrier()
+#define smp_mb__after_clear_bit()	barrier()
+
+static inline void __set_bit(int nr, volatile unsigned long *addr)
+{
+	int *a = (int *)addr;
+	int mask;
+
+	a += nr >> 5;
+	mask = 1 << (nr & 0x1f);
+	*a |= mask;
+}
+
+static inline void __clear_bit(int nr, volatile unsigned long *addr)
+{
+	int *a = (int *)addr;
+	int mask;
+
+	a += nr >> 5;
+	mask = 1 << (nr & 0x1f);
+	*a &= ~mask;
+}
+
+static inline void __change_bit(int nr, volatile unsigned long *addr)
+{
+	int mask;
+	unsigned long *ADDR = (unsigned long *)addr;
+
+	ADDR += nr >> 5;
+	mask = 1 << (nr & 31);
+	*ADDR ^= mask;
+}
+
+static inline int __test_and_set_bit(int nr, volatile unsigned long *addr)
 {
 	int mask, retval;
 	volatile unsigned int *a = (volatile unsigned int *)addr;
@@ -145,26 +212,23 @@ static __inline__ int __test_and_clear_bit(int nr, volatile unsigned long *addr)
 	a += nr >> 5;
 	mask = 1 << (nr & 0x1f);
 	retval = (mask & *a) != 0;
-	*a &= ~mask;
+	*a |= mask;
 	return retval;
 }
 
-static __inline__ int test_and_change_bit(int nr, volatile unsigned long *addr)
+static inline int __test_and_clear_bit(int nr, volatile unsigned long *addr)
 {
 	int mask, retval;
 	volatile unsigned int *a = (volatile unsigned int *)addr;
-	unsigned long flags;
 
 	a += nr >> 5;
 	mask = 1 << (nr & 0x1f);
-	local_irq_save(flags);
 	retval = (mask & *a) != 0;
-	*a ^= mask;
-	local_irq_restore(flags);
+	*a &= ~mask;
 	return retval;
 }
 
-static __inline__ int __test_and_change_bit(int nr,
+static inline int __test_and_change_bit(int nr,
 					    volatile unsigned long *addr)
 {
 	int mask, retval;
@@ -177,16 +241,7 @@ static __inline__ int __test_and_change_bit(int nr,
 	return retval;
 }
 
-/*
- * This routine doesn't need to be atomic.
- */
-static __inline__ int __constant_test_bit(int nr, const void *addr)
-{
-	return ((1UL << (nr & 31)) &
-		(((const volatile unsigned int *)addr)[nr >> 5])) != 0;
-}
-
-static __inline__ int __test_bit(int nr, const void *addr)
+static inline int __test_bit(int nr, const void *addr)
 {
 	int *a = (int *)addr;
 	int mask;
@@ -196,10 +251,16 @@ static __inline__ int __test_bit(int nr, const void *addr)
 	return ((mask & *a) != 0);
 }
 
-#define test_bit(nr,addr) \
-(__builtin_constant_p(nr) ? \
- __constant_test_bit((nr),(addr)) : \
- __test_bit((nr),(addr)))
+#ifndef CONFIG_SMP
+/*
+ * This routine doesn't need irq save and restore ops in UP
+ * context.
+ */
+static inline int test_bit(int nr, const void *addr)
+{
+	return __test_bit(nr, addr);
+}
+#endif
 
 #include <asm-generic/bitops/find.h>
 #include <asm-generic/bitops/hweight.h>
diff --git a/arch/blackfin/include/asm/cache.h b/arch/blackfin/include/asm/cache.h
index 023d72133b5..86637814cf2 100644
--- a/arch/blackfin/include/asm/cache.h
+++ b/arch/blackfin/include/asm/cache.h
@@ -12,6 +12,11 @@
 #define L1_CACHE_BYTES	(1 << L1_CACHE_SHIFT)
 #define SMP_CACHE_BYTES	L1_CACHE_BYTES
 
+#ifdef CONFIG_SMP
+#define __cacheline_aligned
+#else
+#define ____cacheline_aligned
+
 /*
  * Put cacheline_aliged data to L1 data memory
  */
@@ -21,9 +26,33 @@
 		__section__(".data_l1.cacheline_aligned")))
 #endif
 
+#endif
+
 /*
  * largest L1 which this arch supports
  */
 #define L1_CACHE_SHIFT_MAX	5
 
+#if defined(CONFIG_SMP) && \
+    !defined(CONFIG_BFIN_CACHE_COHERENT) && \
+    defined(CONFIG_BFIN_DCACHE)
+#define __ARCH_SYNC_CORE_DCACHE
+#ifndef __ASSEMBLY__
+asmlinkage void __raw_smp_mark_barrier_asm(void);
+asmlinkage void __raw_smp_check_barrier_asm(void);
+
+static inline void smp_mark_barrier(void)
+{
+	__raw_smp_mark_barrier_asm();
+}
+static inline void smp_check_barrier(void)
+{
+	__raw_smp_check_barrier_asm();
+}
+
+void resync_core_dcache(void);
+#endif
+#endif
+
+
 #endif
diff --git a/arch/blackfin/include/asm/cacheflush.h b/arch/blackfin/include/asm/cacheflush.h
index 4403415583f..1b040f5b4fe 100644
--- a/arch/blackfin/include/asm/cacheflush.h
+++ b/arch/blackfin/include/asm/cacheflush.h
@@ -35,6 +35,7 @@ extern void blackfin_icache_flush_range(unsigned long start_address, unsigned lo
 extern void blackfin_dcache_flush_range(unsigned long start_address, unsigned long end_address);
 extern void blackfin_dcache_invalidate_range(unsigned long start_address, unsigned long end_address);
 extern void blackfin_dflush_page(void *page);
+extern void blackfin_invalidate_entire_dcache(void);
 
 #define flush_dcache_mmap_lock(mapping)		do { } while (0)
 #define flush_dcache_mmap_unlock(mapping)	do { } while (0)
@@ -44,12 +45,20 @@ extern void blackfin_dflush_page(void *page);
 #define flush_cache_vmap(start, end)		do { } while (0)
 #define flush_cache_vunmap(start, end)		do { } while (0)
 
+#ifdef CONFIG_SMP
+#define flush_icache_range_others(start, end)	\
+	smp_icache_flush_range_others((start), (end))
+#else
+#define flush_icache_range_others(start, end)	do { } while (0)
+#endif
+
 static inline void flush_icache_range(unsigned start, unsigned end)
 {
 #if defined(CONFIG_BFIN_DCACHE) && defined(CONFIG_BFIN_ICACHE)
 
 # if defined(CONFIG_BFIN_WT)
 	blackfin_icache_flush_range((start), (end));
+	flush_icache_range_others(start, end);
 # else
 	blackfin_icache_dcache_flush_range((start), (end));
 # endif
@@ -58,6 +67,7 @@ static inline void flush_icache_range(unsigned start, unsigned end)
 
 # if defined(CONFIG_BFIN_ICACHE)
 	blackfin_icache_flush_range((start), (end));
+	flush_icache_range_others(start, end);
 # endif
 # if defined(CONFIG_BFIN_DCACHE)
 	blackfin_dcache_flush_range((start), (end));
@@ -66,10 +76,12 @@ static inline void flush_icache_range(unsigned start, unsigned end)
 #endif
 }
 
-#define copy_to_user_page(vma, page, vaddr, dst, src, len) \
-do { memcpy(dst, src, len); \
-     flush_icache_range ((unsigned) (dst), (unsigned) (dst) + (len)); \
+#define copy_to_user_page(vma, page, vaddr, dst, src, len)		\
+do { memcpy(dst, src, len);						\
+     flush_icache_range((unsigned) (dst), (unsigned) (dst) + (len));	\
+     flush_icache_range_others((unsigned long) (dst), (unsigned long) (dst) + (len));\
 } while (0)
+
 #define copy_from_user_page(vma, page, vaddr, dst, src, len)	memcpy(dst, src, len)
 
 #if defined(CONFIG_BFIN_DCACHE)
@@ -82,7 +94,7 @@ do { memcpy(dst, src, len); \
 # define flush_dcache_page(page)			blackfin_dflush_page(page_address(page))
 #else
 # define flush_dcache_range(start,end)		do { } while (0)
-# define flush_dcache_page(page)			do { } while (0)
+# define flush_dcache_page(page)		do { } while (0)
 #endif
 
 extern unsigned long reserved_mem_dcache_on;
diff --git a/arch/blackfin/include/asm/context.S b/arch/blackfin/include/asm/context.S
index c0e630edfb9..40d20b4a9b1 100644
--- a/arch/blackfin/include/asm/context.S
+++ b/arch/blackfin/include/asm/context.S
@@ -303,9 +303,14 @@
 	RETI = [sp++];
 	RETS = [sp++];
 
+#ifdef CONFIG_SMP
+	GET_PDA(p0, r0);
+	r0 = [p0 + PDA_IRQFLAGS];
+#else
 	p0.h = _irq_flags;
 	p0.l = _irq_flags;
 	r0 = [p0];
+#endif
 	sti r0;
 
 	sp += 4;	/* Skip Reserved */
@@ -352,4 +357,3 @@
 	SYSCFG = [sp++];
 	csync;
 .endm
-
diff --git a/arch/blackfin/include/asm/cpu.h b/arch/blackfin/include/asm/cpu.h
new file mode 100644
index 00000000000..9b7aefe7eb2
--- /dev/null
+++ b/arch/blackfin/include/asm/cpu.h
@@ -0,0 +1,42 @@
+/*
+ * File:         arch/blackfin/include/asm/cpu.h.
+ * Author:       Philippe Gerum <rpm@xenomai.org>
+ *
+ *               Copyright 2007 Analog Devices Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#ifndef __ASM_BLACKFIN_CPU_H
+#define __ASM_BLACKFIN_CPU_H
+
+#include <linux/percpu.h>
+
+struct task_struct;
+
+struct blackfin_cpudata {
+	struct cpu cpu;
+	struct task_struct *idle;
+	unsigned long cclk;
+	unsigned int imemctl;
+	unsigned int dmemctl;
+	unsigned long loops_per_jiffy;
+	unsigned long dcache_invld_count;
+};
+
+DECLARE_PER_CPU(struct blackfin_cpudata, cpu_data);
+
+#endif
diff --git a/arch/blackfin/include/asm/l1layout.h b/arch/blackfin/include/asm/l1layout.h
index c13ded77782..06bb37f6c78 100644
--- a/arch/blackfin/include/asm/l1layout.h
+++ b/arch/blackfin/include/asm/l1layout.h
@@ -24,7 +24,8 @@ struct l1_scratch_task_info
 };
 
 /* A pointer to the structure in memory.  */
-#define L1_SCRATCH_TASK_INFO ((struct l1_scratch_task_info *)L1_SCRATCH_START)
+#define L1_SCRATCH_TASK_INFO ((struct l1_scratch_task_info *)\
+						get_l1_scratch_start())
 
 #endif
 
diff --git a/arch/blackfin/include/asm/mutex-dec.h b/arch/blackfin/include/asm/mutex-dec.h
new file mode 100644
index 00000000000..0134151656a
--- /dev/null
+++ b/arch/blackfin/include/asm/mutex-dec.h
@@ -0,0 +1,112 @@
+/*
+ * include/asm-generic/mutex-dec.h
+ *
+ * Generic implementation of the mutex fastpath, based on atomic
+ * decrement/increment.
+ */
+#ifndef _ASM_GENERIC_MUTEX_DEC_H
+#define _ASM_GENERIC_MUTEX_DEC_H
+
+/**
+ *  __mutex_fastpath_lock - try to take the lock by moving the count
+ *                          from 1 to a 0 value
+ *  @count: pointer of type atomic_t
+ *  @fail_fn: function to call if the original value was not 1
+ *
+ * Change the count from 1 to a value lower than 1, and call <fail_fn> if
+ * it wasn't 1 originally. This function MUST leave the value lower than
+ * 1 even when the "1" assertion wasn't true.
+ */
+static inline void
+__mutex_fastpath_lock(atomic_t *count, fastcall void (*fail_fn)(atomic_t *))
+{
+	if (unlikely(atomic_dec_return(count) < 0))
+		fail_fn(count);
+	else
+		smp_mb();
+}
+
+/**
+ *  __mutex_fastpath_lock_retval - try to take the lock by moving the count
+ *                                 from 1 to a 0 value
+ *  @count: pointer of type atomic_t
+ *  @fail_fn: function to call if the original value was not 1
+ *
+ * Change the count from 1 to a value lower than 1, and call <fail_fn> if
+ * it wasn't 1 originally. This function returns 0 if the fastpath succeeds,
+ * or anything the slow path function returns.
+ */
+static inline int
+__mutex_fastpath_lock_retval(atomic_t *count, fastcall int (*fail_fn)(atomic_t *))
+{
+	if (unlikely(atomic_dec_return(count) < 0))
+		return fail_fn(count);
+	else {
+		smp_mb();
+		return 0;
+	}
+}
+
+/**
+ *  __mutex_fastpath_unlock - try to promote the count from 0 to 1
+ *  @count: pointer of type atomic_t
+ *  @fail_fn: function to call if the original value was not 0
+ *
+ * Try to promote the count from 0 to 1. If it wasn't 0, call <fail_fn>.
+ * In the failure case, this function is allowed to either set the value to
+ * 1, or to set it to a value lower than 1.
+ *
+ * If the implementation sets it to a value of lower than 1, then the
+ * __mutex_slowpath_needs_to_unlock() macro needs to return 1, it needs
+ * to return 0 otherwise.
+ */
+static inline void
+__mutex_fastpath_unlock(atomic_t *count, fastcall void (*fail_fn)(atomic_t *))
+{
+	smp_mb();
+	if (unlikely(atomic_inc_return(count) <= 0))
+		fail_fn(count);
+}
+
+#define __mutex_slowpath_needs_to_unlock()		1
+
+/**
+ * __mutex_fastpath_trylock - try to acquire the mutex, without waiting
+ *
+ *  @count: pointer of type atomic_t
+ *  @fail_fn: fallback function
+ *
+ * Change the count from 1 to a value lower than 1, and return 0 (failure)
+ * if it wasn't 1 originally, or return 1 (success) otherwise. This function
+ * MUST leave the value lower than 1 even when the "1" assertion wasn't true.
+ * Additionally, if the value was < 0 originally, this function must not leave
+ * it to 0 on failure.
+ *
+ * If the architecture has no effective trylock variant, it should call the
+ * <fail_fn> spinlock-based trylock variant unconditionally.
+ */
+static inline int
+__mutex_fastpath_trylock(atomic_t *count, int (*fail_fn)(atomic_t *))
+{
+	/*
+	 * We have two variants here. The cmpxchg based one is the best one
+	 * because it never induce a false contention state.  It is included
+	 * here because architectures using the inc/dec algorithms over the
+	 * xchg ones are much more likely to support cmpxchg natively.
+	 *
+	 * If not we fall back to the spinlock based variant - that is
+	 * just as efficient (and simpler) as a 'destructive' probing of
+	 * the mutex state would be.
+	 */
+#ifdef __HAVE_ARCH_CMPXCHG
+	if (likely(atomic_cmpxchg(count, 1, 0) == 1)) {
+		smp_mb();
+		return 1;
+	}
+	return 0;
+#else
+	return fail_fn(count);
+#endif
+}
+
+#endif
diff --git a/arch/blackfin/include/asm/mutex.h b/arch/blackfin/include/asm/mutex.h
index 458c1f7fbc1..5d399256bf0 100644
--- a/arch/blackfin/include/asm/mutex.h
+++ b/arch/blackfin/include/asm/mutex.h
@@ -6,4 +6,67 @@
  * implementation. (see asm-generic/mutex-xchg.h for details)
  */
 
+#ifndef _ASM_MUTEX_H
+#define _ASM_MUTEX_H
+
+#ifndef CONFIG_SMP
 #include <asm-generic/mutex-dec.h>
+#else
+
+static inline void
+__mutex_fastpath_lock(atomic_t *count, void (*fail_fn)(atomic_t *))
+{
+	if (unlikely(atomic_dec_return(count) < 0))
+		fail_fn(count);
+	else
+		smp_mb();
+}
+
+static inline int
+__mutex_fastpath_lock_retval(atomic_t *count, int (*fail_fn)(atomic_t *))
+{
+	if (unlikely(atomic_dec_return(count) < 0))
+		return fail_fn(count);
+	else {
+		smp_mb();
+		return 0;
+	}
+}
+
+static inline void
+__mutex_fastpath_unlock(atomic_t *count, void (*fail_fn)(atomic_t *))
+{
+	smp_mb();
+	if (unlikely(atomic_inc_return(count) <= 0))
+		fail_fn(count);
+}
+
+#define __mutex_slowpath_needs_to_unlock()		1
+
+static inline int
+__mutex_fastpath_trylock(atomic_t *count, int (*fail_fn)(atomic_t *))
+{
+	/*
+	 * We have two variants here. The cmpxchg based one is the best one
+	 * because it never induce a false contention state.  It is included
+	 * here because architectures using the inc/dec algorithms over the
+	 * xchg ones are much more likely to support cmpxchg natively.
+	 *
+	 * If not we fall back to the spinlock based variant - that is
+	 * just as efficient (and simpler) as a 'destructive' probing of
+	 * the mutex state would be.
+	 */
+#ifdef __HAVE_ARCH_CMPXCHG
+	if (likely(atomic_cmpxchg(count, 1, 0) == 1)) {
+		smp_mb();
+		return 1;
+	}
+	return 0;
+#else
+	return fail_fn(count);
+#endif
+}
+
+#endif
+
+#endif
diff --git a/arch/blackfin/include/asm/pda.h b/arch/blackfin/include/asm/pda.h
new file mode 100644
index 00000000000..a24d130c30f
--- /dev/null
+++ b/arch/blackfin/include/asm/pda.h
@@ -0,0 +1,70 @@
+/*
+ * File:         arch/blackfin/include/asm/pda.h
+ * Author:       Philippe Gerum <rpm@xenomai.org>
+ *
+ *               Copyright 2007 Analog Devices Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#ifndef _ASM_BLACKFIN_PDA_H
+#define _ASM_BLACKFIN_PDA_H
+
+#include <asm/mem_map.h>
+
+#ifndef __ASSEMBLY__
+
+struct blackfin_pda {			/* Per-processor Data Area */
+	struct blackfin_pda *next;
+
+	unsigned long syscfg;
+#ifdef CONFIG_SMP
+	unsigned long imask;		/* Current IMASK value */
+#endif
+
+	unsigned long *ipdt;		/* Start of switchable I-CPLB table */
+	unsigned long *ipdt_swapcount;	/* Number of swaps in ipdt */
+	unsigned long *dpdt;		/* Start of switchable D-CPLB table */
+	unsigned long *dpdt_swapcount;	/* Number of swaps in dpdt */
+
+	/*
+	 * Single instructions can have multiple faults, which
+	 * need to be handled by traps.c, in irq5. We store
+	 * the exception cause to ensure we don't miss a
+	 * double fault condition
+	 */
+	unsigned long ex_iptr;
+	unsigned long ex_optr;
+	unsigned long ex_buf[4];
+	unsigned long ex_imask;		/* Saved imask from exception */
+	unsigned long *ex_stack;	/* Exception stack space */
+
+#ifdef ANOMALY_05000261
+	unsigned long last_cplb_fault_retx;
+#endif
+	unsigned long dcplb_fault_addr;
+	unsigned long icplb_fault_addr;
+	unsigned long retx;
+	unsigned long seqstat;
+};
+
+extern struct blackfin_pda cpu_pda[];
+
+void reserve_pda(void);
+
+#endif	/* __ASSEMBLY__ */
+
+#endif /* _ASM_BLACKFIN_PDA_H */
diff --git a/arch/blackfin/include/asm/percpu.h b/arch/blackfin/include/asm/percpu.h
index 78dd61f6b39..797c0c16506 100644
--- a/arch/blackfin/include/asm/percpu.h
+++ b/arch/blackfin/include/asm/percpu.h
@@ -3,4 +3,14 @@
 
 #include <asm-generic/percpu.h>
 
-#endif				/* __ARCH_BLACKFIN_PERCPU__ */
+#ifdef CONFIG_MODULES
+#define PERCPU_MODULE_RESERVE 8192
+#else
+#define PERCPU_MODULE_RESERVE 0
+#endif
+
+#define PERCPU_ENOUGH_ROOM \
+	(ALIGN(__per_cpu_end - __per_cpu_start, SMP_CACHE_BYTES) + \
+	 PERCPU_MODULE_RESERVE)
+
+#endif	/* __ARCH_BLACKFIN_PERCPU__ */
diff --git a/arch/blackfin/include/asm/processor.h b/arch/blackfin/include/asm/processor.h
index e3e9b41fa8d..30703c75030 100644
--- a/arch/blackfin/include/asm/processor.h
+++ b/arch/blackfin/include/asm/processor.h
@@ -106,7 +106,8 @@ unsigned long get_wchan(struct task_struct *p);
 	eip; })
 #define	KSTK_ESP(tsk)	((tsk) == current ? rdusp() : (tsk)->thread.usp)
 
-#define cpu_relax()    	barrier()
+#define cpu_relax()    	smp_mb()
+
 
 /* Get the Silicon Revision of the chip */
 static inline uint32_t __pure bfin_revid(void)
@@ -137,7 +138,11 @@ static inline uint32_t __pure bfin_revid(void)
 static inline uint16_t __pure bfin_cpuid(void)
 {
 	return (bfin_read_CHIPID() & CHIPID_FAMILY) >> 12;
+}
 
+static inline uint32_t __pure bfin_dspid(void)
+{
+	return bfin_read_DSPID();
 }
 
 static inline uint32_t __pure bfin_compiled_revid(void)
diff --git a/arch/blackfin/include/asm/rwlock.h b/arch/blackfin/include/asm/rwlock.h
new file mode 100644
index 00000000000..4a724b37897
--- /dev/null
+++ b/arch/blackfin/include/asm/rwlock.h
@@ -0,0 +1,6 @@
+#ifndef _ASM_BLACKFIN_RWLOCK_H
+#define _ASM_BLACKFIN_RWLOCK_H
+
+#define RW_LOCK_BIAS	0x01000000
+
+#endif
diff --git a/arch/blackfin/include/asm/smp.h b/arch/blackfin/include/asm/smp.h
new file mode 100644
index 00000000000..233cb8c3cfb
--- /dev/null
+++ b/arch/blackfin/include/asm/smp.h
@@ -0,0 +1,42 @@
+/*
+ * File:         arch/blackfin/include/asm/smp.h
+ * Author:       Philippe Gerum <rpm@xenomai.org>
+ *
+ *               Copyright 2007 Analog Devices Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#ifndef __ASM_BLACKFIN_SMP_H
+#define __ASM_BLACKFIN_SMP_H
+
+#include <linux/kernel.h>
+#include <linux/threads.h>
+#include <linux/cpumask.h>
+#include <linux/cache.h>
+#include <asm/blackfin.h>
+#include <mach/smp.h>
+
+#define raw_smp_processor_id()  blackfin_core_id()
+
+struct corelock_slot {
+	int lock;
+};
+
+void smp_icache_flush_range_others(unsigned long start,
+				   unsigned long end);
+
+#endif /* !__ASM_BLACKFIN_SMP_H */
diff --git a/arch/blackfin/include/asm/spinlock.h b/arch/blackfin/include/asm/spinlock.h
index 64e908a5064..0249ac31947 100644
--- a/arch/blackfin/include/asm/spinlock.h
+++ b/arch/blackfin/include/asm/spinlock.h
@@ -1,6 +1,89 @@
 #ifndef __BFIN_SPINLOCK_H
 #define __BFIN_SPINLOCK_H
 
-#error blackfin architecture does not support SMP spin lock yet
+#include <asm/atomic.h>
 
-#endif
+asmlinkage int __raw_spin_is_locked_asm(volatile int *ptr);
+asmlinkage void __raw_spin_lock_asm(volatile int *ptr);
+asmlinkage int __raw_spin_trylock_asm(volatile int *ptr);
+asmlinkage void __raw_spin_unlock_asm(volatile int *ptr);
+asmlinkage void __raw_read_lock_asm(volatile int *ptr);
+asmlinkage int __raw_read_trylock_asm(volatile int *ptr);
+asmlinkage void __raw_read_unlock_asm(volatile int *ptr);
+asmlinkage void __raw_write_lock_asm(volatile int *ptr);
+asmlinkage int __raw_write_trylock_asm(volatile int *ptr);
+asmlinkage void __raw_write_unlock_asm(volatile int *ptr);
+
+static inline int __raw_spin_is_locked(raw_spinlock_t *lock)
+{
+	return __raw_spin_is_locked_asm(&lock->lock);
+}
+
+static inline void __raw_spin_lock(raw_spinlock_t *lock)
+{
+	__raw_spin_lock_asm(&lock->lock);
+}
+
+#define __raw_spin_lock_flags(lock, flags) __raw_spin_lock(lock)
+
+static inline int __raw_spin_trylock(raw_spinlock_t *lock)
+{
+	return __raw_spin_trylock_asm(&lock->lock);
+}
+
+static inline void __raw_spin_unlock(raw_spinlock_t *lock)
+{
+	__raw_spin_unlock_asm(&lock->lock);
+}
+
+static inline void __raw_spin_unlock_wait(raw_spinlock_t *lock)
+{
+	while (__raw_spin_is_locked(lock))
+		cpu_relax();
+}
+
+static inline int __raw_read_can_lock(raw_rwlock_t *rw)
+{
+	return __raw_uncached_fetch_asm(&rw->lock) > 0;
+}
+
+static inline int __raw_write_can_lock(raw_rwlock_t *rw)
+{
+	return __raw_uncached_fetch_asm(&rw->lock) == RW_LOCK_BIAS;
+}
+
+static inline void __raw_read_lock(raw_rwlock_t *rw)
+{
+	__raw_read_lock_asm(&rw->lock);
+}
+
+static inline int __raw_read_trylock(raw_rwlock_t *rw)
+{
+	return __raw_read_trylock_asm(&rw->lock);
+}
+
+static inline void __raw_read_unlock(raw_rwlock_t *rw)
+{
+	__raw_read_unlock_asm(&rw->lock);
+}
+
+static inline void __raw_write_lock(raw_rwlock_t *rw)
+{
+	__raw_write_lock_asm(&rw->lock);
+}
+
+static inline int __raw_write_trylock(raw_rwlock_t *rw)
+{
+	return __raw_write_trylock_asm(&rw->lock);
+}
+
+static inline void __raw_write_unlock(raw_rwlock_t *rw)
+{
+	__raw_write_unlock_asm(&rw->lock);
+}
+
+#define _raw_spin_relax(lock)  	cpu_relax()
+#define _raw_read_relax(lock)	cpu_relax()
+#define _raw_write_relax(lock)	cpu_relax()
+
+#endif /*  !__BFIN_SPINLOCK_H */
diff --git a/arch/blackfin/include/asm/spinlock_types.h b/arch/blackfin/include/asm/spinlock_types.h
new file mode 100644
index 00000000000..b1e3c4c7b38
--- /dev/null
+++ b/arch/blackfin/include/asm/spinlock_types.h
@@ -0,0 +1,22 @@
+#ifndef __ASM_SPINLOCK_TYPES_H
+#define __ASM_SPINLOCK_TYPES_H
+
+#ifndef __LINUX_SPINLOCK_TYPES_H
+# error "please don't include this file directly"
+#endif
+
+#include <asm/rwlock.h>
+
+typedef struct {
+	volatile unsigned int lock;
+} raw_spinlock_t;
+
+#define __RAW_SPIN_LOCK_UNLOCKED	{ 0 }
+
+typedef struct {
+	volatile unsigned int lock;
+} raw_rwlock_t;
+
+#define __RAW_RW_LOCK_UNLOCKED		{ RW_LOCK_BIAS }
+
+#endif
diff --git a/arch/blackfin/include/asm/system.h b/arch/blackfin/include/asm/system.h
index 8f1627d8bf0..6b368faf30c 100644
--- a/arch/blackfin/include/asm/system.h
+++ b/arch/blackfin/include/asm/system.h
@@ -37,20 +37,16 @@
 #include <linux/linkage.h>
 #include <linux/compiler.h>
 #include <mach/anomaly.h>
+#include <asm/pda.h>
+#include <asm/processor.h>
+
+/* Forward decl needed due to cdef inter dependencies */
+static inline uint32_t __pure bfin_dspid(void);
+#define blackfin_core_id() (bfin_dspid() & 0xff)
 
 /*
  * Interrupt configuring macros.
  */
-
-extern unsigned long irq_flags;
-
-#define local_irq_enable() \
-	__asm__ __volatile__( \
-		"sti %0;" \
-		: \
-		: "d" (irq_flags) \
-	)
-
 #define local_irq_disable() \
 	do { \
 		int __tmp_dummy; \
@@ -66,6 +62,18 @@ extern unsigned long irq_flags;
 # define NOP_PAD_ANOMALY_05000244
 #endif
 
+#ifdef CONFIG_SMP
+# define irq_flags cpu_pda[blackfin_core_id()].imask
+#else
+extern unsigned long irq_flags;
+#endif
+
+#define local_irq_enable() \
+	__asm__ __volatile__( \
+		"sti %0;" \
+		: \
+		: "d" (irq_flags) \
+	)
 #define idle_with_irq_disabled() \
 	__asm__ __volatile__( \
 		NOP_PAD_ANOMALY_05000244 \
@@ -129,22 +137,85 @@ extern unsigned long irq_flags;
 #define rmb()  asm volatile (""   : : :"memory")
 #define wmb()  asm volatile (""   : : :"memory")
 #define set_mb(var, value) do { (void) xchg(&var, value); } while (0)
-
 #define read_barrier_depends() 		do { } while(0)
 
 #ifdef CONFIG_SMP
-#define smp_mb()	mb()
-#define smp_rmb()	rmb()
-#define smp_wmb()	wmb()
-#define smp_read_barrier_depends()	read_barrier_depends()
+asmlinkage unsigned long __raw_xchg_1_asm(volatile void *ptr, unsigned long value);
+asmlinkage unsigned long __raw_xchg_2_asm(volatile void *ptr, unsigned long value);
+asmlinkage unsigned long __raw_xchg_4_asm(volatile void *ptr, unsigned long value);
+asmlinkage unsigned long __raw_cmpxchg_1_asm(volatile void *ptr,
+					unsigned long new, unsigned long old);
+asmlinkage unsigned long __raw_cmpxchg_2_asm(volatile void *ptr,
+					unsigned long new, unsigned long old);
+asmlinkage unsigned long __raw_cmpxchg_4_asm(volatile void *ptr,
+					unsigned long new, unsigned long old);
+
+#ifdef __ARCH_SYNC_CORE_DCACHE
+# define smp_mb()	do { barrier(); smp_check_barrier(); smp_mark_barrier(); } while (0)
+# define smp_rmb()	do { barrier(); smp_check_barrier(); } while (0)
+# define smp_wmb()	do { barrier(); smp_mark_barrier(); } while (0)
 #else
+# define smp_mb()	barrier()
+# define smp_rmb()	barrier()
+# define smp_wmb()	barrier()
+#endif
+
+static inline unsigned long __xchg(unsigned long x, volatile void *ptr,
+				   int size)
+{
+	unsigned long tmp;
+
+	switch (size) {
+	case 1:
+		tmp = __raw_xchg_1_asm(ptr, x);
+		break;
+	case 2:
+		tmp = __raw_xchg_2_asm(ptr, x);
+		break;
+	case 4:
+		tmp = __raw_xchg_4_asm(ptr, x);
+		break;
+	}
+
+	return tmp;
+}
+
+/*
+ * Atomic compare and exchange.  Compare OLD with MEM, if identical,
+ * store NEW in MEM.  Return the initial value in MEM.  Success is
+ * indicated by comparing RETURN with OLD.
+ */
+static inline unsigned long __cmpxchg(volatile void *ptr, unsigned long old,
+				      unsigned long new, int size)
+{
+	unsigned long tmp;
+
+	switch (size) {
+	case 1:
+		tmp = __raw_cmpxchg_1_asm(ptr, new, old);
+		break;
+	case 2:
+		tmp = __raw_cmpxchg_2_asm(ptr, new, old);
+		break;
+	case 4:
+		tmp = __raw_cmpxchg_4_asm(ptr, new, old);
+		break;
+	}
+
+	return tmp;
+}
+#define cmpxchg(ptr, o, n) \
+	((__typeof__(*(ptr)))__cmpxchg((ptr), (unsigned long)(o), \
+		(unsigned long)(n), sizeof(*(ptr))))
+
+#define smp_read_barrier_depends()	smp_check_barrier()
+
+#else /* !CONFIG_SMP */
+
 #define smp_mb()	barrier()
 #define smp_rmb()	barrier()
 #define smp_wmb()	barrier()
 #define smp_read_barrier_depends()	do { } while(0)
-#endif
-
-#define xchg(ptr,x) ((__typeof__(*(ptr)))__xchg((unsigned long)(x),(ptr),sizeof(*(ptr))))
 
 struct __xchg_dummy {
 	unsigned long a[100];
@@ -194,9 +265,12 @@ static inline unsigned long __xchg(unsigned long x, volatile void *ptr,
 			(unsigned long)(n), sizeof(*(ptr))))
 #define cmpxchg64_local(ptr, o, n) __cmpxchg64_local_generic((ptr), (o), (n))
 
-#ifndef CONFIG_SMP
 #include <asm-generic/cmpxchg.h>
-#endif
+
+#endif /* !CONFIG_SMP */
+
+#define xchg(ptr, x) ((__typeof__(*(ptr)))__xchg((unsigned long)(x), (ptr), sizeof(*(ptr))))
+#define tas(ptr) ((void)xchg((ptr), 1))
 
 #define prepare_to_switch()     do { } while(0)
 
@@ -218,4 +292,4 @@ do {    \
 	(last) = resume (prev, next);   \
 } while (0)
 
-#endif				/* _BLACKFIN_SYSTEM_H */
+#endif	/* _BLACKFIN_SYSTEM_H */