MIPS: Get rid of branches to .subsections.

It was a nice optimization - on paper at least.  In practice it results in
branches that may exceed the maximum legal range for a branch.  We can
fight that problem with -ffunction-sections but -ffunction-sections again
is incompatible with -pg used by the function tracer.

By rewriting the loop around all simple LL/SC blocks to C we reduce the
amount of inline assembler and at the same time allow GCC to often fill
the branch delay slots with something sensible or whatever else clever
optimization it may have up in its sleeve.

With this optimization gone we also no longer need -ffunction-sections,
so drop it.

This optimization was originally introduced in 2.6.21, commit
5999eca25c1fd4b9b9aca7833b04d10fe4bc877d (linux-mips.org) rsp.
f65e4fa8e0c6022ad58dc88d1b11b12589ed7f9f (kernel.org).

Original fix for the issues which caused me to pull this optimization by
Paul Gortmaker <paul.gortmaker@windriver.com>.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

1 files changed, 96 insertions, 112 deletions

diff --git a/arch/mips/include/asm/atomic.h b/arch/mips/include/asm/atomic.h
index 47d87da379f9..4a02fe891ab6 100644
--- a/arch/mips/include/asm/atomic.h
+++ b/arch/mips/include/asm/atomic.h
@@ -64,18 +64,16 @@ static __inline__ void atomic_add(int i, atomic_t * v)
 	} else if (kernel_uses_llsc) {
 		int temp;
 
-		__asm__ __volatile__(
-		"	.set	mips3					\n"
-		"1:	ll	%0, %1		# atomic_add		\n"
-		"	addu	%0, %2					\n"
-		"	sc	%0, %1					\n"
-		"	beqz	%0, 2f					\n"
-		"	.subsection 2					\n"
-		"2:	b	1b					\n"
-		"	.previous					\n"
-		"	.set	mips0					\n"
-		: "=&r" (temp), "=m" (v->counter)
-		: "Ir" (i), "m" (v->counter));
+		do {
+			__asm__ __volatile__(
+			"	.set	mips3				\n"
+			"	ll	%0, %1		# atomic_add	\n"
+			"	addu	%0, %2				\n"
+			"	sc	%0, %1				\n"
+			"	.set	mips0				\n"
+			: "=&r" (temp), "=m" (v->counter)
+			: "Ir" (i), "m" (v->counter));
+		} while (unlikely(!temp));
 	} else {
 		unsigned long flags;
 
@@ -109,18 +107,16 @@ static __inline__ void atomic_sub(int i, atomic_t * v)
 	} else if (kernel_uses_llsc) {
 		int temp;
 
-		__asm__ __volatile__(
-		"	.set	mips3					\n"
-		"1:	ll	%0, %1		# atomic_sub		\n"
-		"	subu	%0, %2					\n"
-		"	sc	%0, %1					\n"
-		"	beqz	%0, 2f					\n"
-		"	.subsection 2					\n"
-		"2:	b	1b					\n"
-		"	.previous					\n"
-		"	.set	mips0					\n"
-		: "=&r" (temp), "=m" (v->counter)
-		: "Ir" (i), "m" (v->counter));
+		do {
+			__asm__ __volatile__(
+			"	.set	mips3				\n"
+			"	ll	%0, %1		# atomic_sub	\n"
+			"	subu	%0, %2				\n"
+			"	sc	%0, %1				\n"
+			"	.set	mips0				\n"
+			: "=&r" (temp), "=m" (v->counter)
+			: "Ir" (i), "m" (v->counter));
+		} while (unlikely(!temp));
 	} else {
 		unsigned long flags;
 
@@ -156,20 +152,19 @@ static __inline__ int atomic_add_return(int i, atomic_t * v)
 	} else if (kernel_uses_llsc) {
 		int temp;
 
-		__asm__ __volatile__(
-		"	.set	mips3					\n"
-		"1:	ll	%1, %2		# atomic_add_return	\n"
-		"	addu	%0, %1, %3				\n"
-		"	sc	%0, %2					\n"
-		"	beqz	%0, 2f					\n"
-		"	addu	%0, %1, %3				\n"
-		"	.subsection 2					\n"
-		"2:	b	1b					\n"
-		"	.previous					\n"
-		"	.set	mips0					\n"
-		: "=&r" (result), "=&r" (temp), "=m" (v->counter)
-		: "Ir" (i), "m" (v->counter)
-		: "memory");
+		do {
+			__asm__ __volatile__(
+			"	.set	mips3				\n"
+			"	ll	%1, %2	# atomic_add_return	\n"
+			"	addu	%0, %1, %3			\n"
+			"	sc	%0, %2				\n"
+			"	.set	mips0				\n"
+			: "=&r" (result), "=&r" (temp), "=m" (v->counter)
+			: "Ir" (i), "m" (v->counter)
+			: "memory");
+		} while (unlikely(!result));
+
+		result = temp + i;
 	} else {
 		unsigned long flags;
 
@@ -205,23 +200,24 @@ static __inline__ int atomic_sub_return(int i, atomic_t * v)
 		: "=&r" (result), "=&r" (temp), "=m" (v->counter)
 		: "Ir" (i), "m" (v->counter)
 		: "memory");
+
+		result = temp - i;
 	} else if (kernel_uses_llsc) {
 		int temp;
 
-		__asm__ __volatile__(
-		"	.set	mips3					\n"
-		"1:	ll	%1, %2		# atomic_sub_return	\n"
-		"	subu	%0, %1, %3				\n"
-		"	sc	%0, %2					\n"
-		"	beqz	%0, 2f					\n"
-		"	subu	%0, %1, %3				\n"
-		"	.subsection 2					\n"
-		"2:	b	1b					\n"
-		"	.previous					\n"
-		"	.set	mips0					\n"
-		: "=&r" (result), "=&r" (temp), "=m" (v->counter)
-		: "Ir" (i), "m" (v->counter)
-		: "memory");
+		do {
+			__asm__ __volatile__(
+			"	.set	mips3				\n"
+			"	ll	%1, %2	# atomic_sub_return	\n"
+			"	subu	%0, %1, %3			\n"
+			"	sc	%0, %2				\n"
+			"	.set	mips0				\n"
+			: "=&r" (result), "=&r" (temp), "=m" (v->counter)
+			: "Ir" (i), "m" (v->counter)
+			: "memory");
+		} while (unlikely(!result));
+
+		result = temp - i;
 	} else {
 		unsigned long flags;
 
@@ -279,12 +275,9 @@ static __inline__ int atomic_sub_if_positive(int i, atomic_t * v)
 		"	bltz	%0, 1f					\n"
 		"	sc	%0, %2					\n"
 		"	.set	noreorder				\n"
-		"	beqz	%0, 2f					\n"
+		"	beqz	%0, 1b					\n"
 		"	 subu	%0, %1, %3				\n"
 		"	.set	reorder					\n"
-		"	.subsection 2					\n"
-		"2:	b	1b					\n"
-		"	.previous					\n"
 		"1:							\n"
 		"	.set	mips0					\n"
 		: "=&r" (result), "=&r" (temp), "=m" (v->counter)
@@ -443,18 +436,16 @@ static __inline__ void atomic64_add(long i, atomic64_t * v)
 	} else if (kernel_uses_llsc) {
 		long temp;
 
-		__asm__ __volatile__(
-		"	.set	mips3					\n"
-		"1:	lld	%0, %1		# atomic64_add		\n"
-		"	daddu	%0, %2					\n"
-		"	scd	%0, %1					\n"
-		"	beqz	%0, 2f					\n"
-		"	.subsection 2					\n"
-		"2:	b	1b					\n"
-		"	.previous					\n"
-		"	.set	mips0					\n"
-		: "=&r" (temp), "=m" (v->counter)
-		: "Ir" (i), "m" (v->counter));
+		do {
+			__asm__ __volatile__(
+			"	.set	mips3				\n"
+			"	lld	%0, %1		# atomic64_add	\n"
+			"	daddu	%0, %2				\n"
+			"	scd	%0, %1				\n"
+			"	.set	mips0				\n"
+			: "=&r" (temp), "=m" (v->counter)
+			: "Ir" (i), "m" (v->counter));
+		} while (unlikely(!temp));
 	} else {
 		unsigned long flags;
 
@@ -488,18 +479,16 @@ static __inline__ void atomic64_sub(long i, atomic64_t * v)
 	} else if (kernel_uses_llsc) {
 		long temp;
 
-		__asm__ __volatile__(
-		"	.set	mips3					\n"
-		"1:	lld	%0, %1		# atomic64_sub		\n"
-		"	dsubu	%0, %2					\n"
-		"	scd	%0, %1					\n"
-		"	beqz	%0, 2f					\n"
-		"	.subsection 2					\n"
-		"2:	b	1b					\n"
-		"	.previous					\n"
-		"	.set	mips0					\n"
-		: "=&r" (temp), "=m" (v->counter)
-		: "Ir" (i), "m" (v->counter));
+		do {
+			__asm__ __volatile__(
+			"	.set	mips3				\n"
+			"	lld	%0, %1		# atomic64_sub	\n"
+			"	dsubu	%0, %2				\n"
+			"	scd	%0, %1				\n"
+			"	.set	mips0				\n"
+			: "=&r" (temp), "=m" (v->counter)
+			: "Ir" (i), "m" (v->counter));
+		} while (unlikely(!temp));
 	} else {
 		unsigned long flags;
 
@@ -535,20 +524,19 @@ static __inline__ long atomic64_add_return(long i, atomic64_t * v)
 	} else if (kernel_uses_llsc) {
 		long temp;
 
-		__asm__ __volatile__(
-		"	.set	mips3					\n"
-		"1:	lld	%1, %2		# atomic64_add_return	\n"
-		"	daddu	%0, %1, %3				\n"
-		"	scd	%0, %2					\n"
-		"	beqz	%0, 2f					\n"
-		"	daddu	%0, %1, %3				\n"
-		"	.subsection 2					\n"
-		"2:	b	1b					\n"
-		"	.previous					\n"
-		"	.set	mips0					\n"
-		: "=&r" (result), "=&r" (temp), "=m" (v->counter)
-		: "Ir" (i), "m" (v->counter)
-		: "memory");
+		do {
+			__asm__ __volatile__(
+			"	.set	mips3				\n"
+			"	lld	%1, %2	# atomic64_add_return	\n"
+			"	daddu	%0, %1, %3			\n"
+			"	scd	%0, %2				\n"
+			"	.set	mips0				\n"
+			: "=&r" (result), "=&r" (temp), "=m" (v->counter)
+			: "Ir" (i), "m" (v->counter)
+			: "memory");
+		} while (unlikely(!result));
+
+		result = temp + i;
 	} else {
 		unsigned long flags;
 
@@ -587,20 +575,19 @@ static __inline__ long atomic64_sub_return(long i, atomic64_t * v)
 	} else if (kernel_uses_llsc) {
 		long temp;
 
-		__asm__ __volatile__(
-		"	.set	mips3					\n"
-		"1:	lld	%1, %2		# atomic64_sub_return	\n"
-		"	dsubu	%0, %1, %3				\n"
-		"	scd	%0, %2					\n"
-		"	beqz	%0, 2f					\n"
-		"	dsubu	%0, %1, %3				\n"
-		"	.subsection 2					\n"
-		"2:	b	1b					\n"
-		"	.previous					\n"
-		"	.set	mips0					\n"
-		: "=&r" (result), "=&r" (temp), "=m" (v->counter)
-		: "Ir" (i), "m" (v->counter)
-		: "memory");
+		do {
+			__asm__ __volatile__(
+			"	.set	mips3				\n"
+			"	lld	%1, %2	# atomic64_sub_return	\n"
+			"	dsubu	%0, %1, %3			\n"
+			"	scd	%0, %2				\n"
+			"	.set	mips0				\n"
+			: "=&r" (result), "=&r" (temp), "=m" (v->counter)
+			: "Ir" (i), "m" (v->counter)
+			: "memory");
+		} while (unlikely(!result));
+
+		result = temp - i;
 	} else {
 		unsigned long flags;
 
@@ -658,12 +645,9 @@ static __inline__ long atomic64_sub_if_positive(long i, atomic64_t * v)
 		"	bltz	%0, 1f					\n"
 		"	scd	%0, %2					\n"
 		"	.set	noreorder				\n"
-		"	beqz	%0, 2f					\n"
+		"	beqz	%0, 1b					\n"
 		"	 dsubu	%0, %1, %3				\n"
 		"	.set	reorder					\n"
-		"	.subsection 2					\n"
-		"2:	b	1b					\n"
-		"	.previous					\n"
 		"1:							\n"
 		"	.set	mips0					\n"
 		: "=&r" (result), "=&r" (temp), "=m" (v->counter)