diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index c8fa1443e2c1..49778bb43782 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -14,6 +14,7 @@ config ARM
 	select HAVE_FUNCTION_TRACER if (!XIP_KERNEL)
 	select HAVE_FTRACE_MCOUNT_RECORD if (!XIP_KERNEL)
 	select HAVE_DYNAMIC_FTRACE if (!XIP_KERNEL)
+	select HAVE_FUNCTION_GRAPH_TRACER if (!THUMB2_KERNEL)
 	select HAVE_GENERIC_DMA_COHERENT
 	select HAVE_KERNEL_GZIP
 	select HAVE_KERNEL_LZO
diff --git a/arch/arm/Kconfig.debug b/arch/arm/Kconfig.debug
index 2fd0b99afc4b..eac62085f5b2 100644
--- a/arch/arm/Kconfig.debug
+++ b/arch/arm/Kconfig.debug
@@ -23,7 +23,7 @@ config STRICT_DEVMEM
 config FRAME_POINTER
 	bool
 	depends on !THUMB2_KERNEL
-	default y if !ARM_UNWIND
+	default y if !ARM_UNWIND || FUNCTION_GRAPH_TRACER
 	help
 	  If you say N here, the resulting kernel will be slightly smaller and
 	  faster. However, if neither FRAME_POINTER nor ARM_UNWIND are enabled,
diff --git a/arch/arm/include/asm/system.h b/arch/arm/include/asm/system.h
index 1120f18a6b17..ec4327a4653d 100644
--- a/arch/arm/include/asm/system.h
+++ b/arch/arm/include/asm/system.h
@@ -63,6 +63,11 @@
 #include <asm/outercache.h>
 
 #define __exception	__attribute__((section(".exception.text")))
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+#define __exception_irq_entry	__irq_entry
+#else
+#define __exception_irq_entry	__exception
+#endif
 
 struct thread_info;
 struct task_struct;
diff --git a/arch/arm/include/asm/traps.h b/arch/arm/include/asm/traps.h
index 491960bf4260..124475afb007 100644
--- a/arch/arm/include/asm/traps.h
+++ b/arch/arm/include/asm/traps.h
@@ -15,13 +15,32 @@ struct undef_hook {
 void register_undef_hook(struct undef_hook *hook);
 void unregister_undef_hook(struct undef_hook *hook);
 
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+static inline int __in_irqentry_text(unsigned long ptr)
+{
+	extern char __irqentry_text_start[];
+	extern char __irqentry_text_end[];
+
+	return ptr >= (unsigned long)&__irqentry_text_start &&
+	       ptr < (unsigned long)&__irqentry_text_end;
+}
+#else
+static inline int __in_irqentry_text(unsigned long ptr)
+{
+	return 0;
+}
+#endif
+
 static inline int in_exception_text(unsigned long ptr)
 {
 	extern char __exception_text_start[];
 	extern char __exception_text_end[];
+	int in;
 
-	return ptr >= (unsigned long)&__exception_text_start &&
-	       ptr < (unsigned long)&__exception_text_end;
+	in = ptr >= (unsigned long)&__exception_text_start &&
+	     ptr < (unsigned long)&__exception_text_end;
+
+	return in ? : __in_irqentry_text(ptr);
 }
 
 extern void __init early_trap_init(void);
diff --git a/arch/arm/kernel/Makefile b/arch/arm/kernel/Makefile
index 5b9b268f4fbb..679851a9f589 100644
--- a/arch/arm/kernel/Makefile
+++ b/arch/arm/kernel/Makefile
@@ -5,7 +5,7 @@
 CPPFLAGS_vmlinux.lds := -DTEXT_OFFSET=$(TEXT_OFFSET)
 AFLAGS_head.o        := -DTEXT_OFFSET=$(TEXT_OFFSET)
 
-ifdef CONFIG_DYNAMIC_FTRACE
+ifdef CONFIG_FUNCTION_TRACER
 CFLAGS_REMOVE_ftrace.o = -pg
 endif
 
@@ -33,6 +33,7 @@ obj-$(CONFIG_SMP)		+= smp.o
 obj-$(CONFIG_HAVE_ARM_SCU)	+= smp_scu.o
 obj-$(CONFIG_HAVE_ARM_TWD)	+= smp_twd.o
 obj-$(CONFIG_DYNAMIC_FTRACE)	+= ftrace.o
+obj-$(CONFIG_FUNCTION_GRAPH_TRACER)	+= ftrace.o
 obj-$(CONFIG_KEXEC)		+= machine_kexec.o relocate_kernel.o
 obj-$(CONFIG_KPROBES)		+= kprobes.o kprobes-decode.o
 obj-$(CONFIG_ATAGS_PROC)	+= atags.o
diff --git a/arch/arm/kernel/entry-common.S b/arch/arm/kernel/entry-common.S
index 8bfa98757cd2..aae802ee12f8 100644
--- a/arch/arm/kernel/entry-common.S
+++ b/arch/arm/kernel/entry-common.S
@@ -141,98 +141,170 @@ ENDPROC(ret_from_fork)
 #endif
 #endif
 
-#ifdef CONFIG_DYNAMIC_FTRACE
-ENTRY(__gnu_mcount_nc)
-	mov	ip, lr
-	ldmia	sp!, {lr}
-	mov	pc, ip
-ENDPROC(__gnu_mcount_nc)
-
-ENTRY(ftrace_caller)
-	stmdb	sp!, {r0-r3, lr}
-	mov	r0, lr
-	sub	r0, r0, #MCOUNT_INSN_SIZE
-	ldr	r1, [sp, #20]
-
-	.global	ftrace_call
-ftrace_call:
-	bl	ftrace_stub
-	ldmia	sp!, {r0-r3, ip, lr}
-	mov	pc, ip
-ENDPROC(ftrace_caller)
-
-#ifdef CONFIG_OLD_MCOUNT
-ENTRY(mcount)
-	stmdb	sp!, {lr}
-	ldr	lr, [fp, #-4]
-	ldmia	sp!, {pc}
-ENDPROC(mcount)
-
-ENTRY(ftrace_caller_old)
-	stmdb	sp!, {r0-r3, lr}
-	ldr	r1, [fp, #-4]
-	mov	r0, lr
-	sub	r0, r0, #MCOUNT_INSN_SIZE
-
-	.globl ftrace_call_old
-ftrace_call_old:
-	bl	ftrace_stub
-	ldr	lr, [fp, #-4]			@ restore lr
-	ldmia	sp!, {r0-r3, pc}
-ENDPROC(ftrace_caller_old)
-#endif
-
-#else
-
-ENTRY(__gnu_mcount_nc)
-	stmdb	sp!, {r0-r3, lr}
+.macro __mcount suffix
+	mcount_enter
 	ldr	r0, =ftrace_trace_function
 	ldr	r2, [r0]
 	adr	r0, .Lftrace_stub
 	cmp	r0, r2
-	bne	gnu_trace
-	ldmia	sp!, {r0-r3, ip, lr}
-	mov	pc, ip
+	bne	1f
 
-gnu_trace:
-	ldr	r1, [sp, #20]			@ lr of instrumented routine
-	mov	r0, lr
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+	ldr     r1, =ftrace_graph_return
+	ldr     r2, [r1]
+	cmp     r0, r2
+	bne     ftrace_graph_caller\suffix
+
+	ldr     r1, =ftrace_graph_entry
+	ldr     r2, [r1]
+	ldr     r0, =ftrace_graph_entry_stub
+	cmp     r0, r2
+	bne     ftrace_graph_caller\suffix
+#endif
+
+	mcount_exit
+
+1: 	mcount_get_lr	r1			@ lr of instrumented func
+	mov	r0, lr				@ instrumented function
 	sub	r0, r0, #MCOUNT_INSN_SIZE
-	adr	lr, BSYM(1f)
+	adr	lr, BSYM(2f)
 	mov	pc, r2
-1:
-	ldmia	sp!, {r0-r3, ip, lr}
-	mov	pc, ip
-ENDPROC(__gnu_mcount_nc)
+2:	mcount_exit
+.endm
+
+.macro __ftrace_caller suffix
+	mcount_enter
+
+	mcount_get_lr	r1			@ lr of instrumented func
+	mov	r0, lr				@ instrumented function
+	sub	r0, r0, #MCOUNT_INSN_SIZE
+
+	.globl ftrace_call\suffix
+ftrace_call\suffix:
+	bl	ftrace_stub
+
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+	.globl ftrace_graph_call\suffix
+ftrace_graph_call\suffix:
+	mov	r0, r0
+#endif
+
+	mcount_exit
+.endm
+
+.macro __ftrace_graph_caller
+	sub	r0, fp, #4		@ &lr of instrumented routine (&parent)
+#ifdef CONFIG_DYNAMIC_FTRACE
+	@ called from __ftrace_caller, saved in mcount_enter
+	ldr	r1, [sp, #16]		@ instrumented routine (func)
+#else
+	@ called from __mcount, untouched in lr
+	mov	r1, lr			@ instrumented routine (func)
+#endif
+	sub	r1, r1, #MCOUNT_INSN_SIZE
+	mov	r2, fp			@ frame pointer
+	bl	prepare_ftrace_return
+	mcount_exit
+.endm
 
 #ifdef CONFIG_OLD_MCOUNT
 /*
- * This is under an ifdef in order to force link-time errors for people trying
- * to build with !FRAME_POINTER with a GCC which doesn't use the new-style
- * mcount.
+ * mcount
  */
-ENTRY(mcount)
-	stmdb	sp!, {r0-r3, lr}
-	ldr	r0, =ftrace_trace_function
-	ldr	r2, [r0]
-	adr	r0, ftrace_stub
-	cmp	r0, r2
-	bne	trace
-	ldr	lr, [fp, #-4]			@ restore lr
-	ldmia	sp!, {r0-r3, pc}
 
-trace:
-	ldr	r1, [fp, #-4]			@ lr of instrumented routine
-	mov	r0, lr
-	sub	r0, r0, #MCOUNT_INSN_SIZE
-	mov	lr, pc
-	mov	pc, r2
-	ldr	lr, [fp, #-4]			@ restore lr
+.macro mcount_enter
+	stmdb	sp!, {r0-r3, lr}
+.endm
+
+.macro mcount_get_lr reg
+	ldr	\reg, [fp, #-4]
+.endm
+
+.macro mcount_exit
+	ldr	lr, [fp, #-4]
 	ldmia	sp!, {r0-r3, pc}
+.endm
+
+ENTRY(mcount)
+#ifdef CONFIG_DYNAMIC_FTRACE
+	stmdb	sp!, {lr}
+	ldr	lr, [fp, #-4]
+	ldmia	sp!, {pc}
+#else
+	__mcount _old
+#endif
 ENDPROC(mcount)
+
+#ifdef CONFIG_DYNAMIC_FTRACE
+ENTRY(ftrace_caller_old)
+	__ftrace_caller _old
+ENDPROC(ftrace_caller_old)
 #endif
 
-#endif /* CONFIG_DYNAMIC_FTRACE */
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+ENTRY(ftrace_graph_caller_old)
+	__ftrace_graph_caller
+ENDPROC(ftrace_graph_caller_old)
+#endif
+
+.purgem mcount_enter
+.purgem mcount_get_lr
+.purgem mcount_exit
+#endif
+
+/*
+ * __gnu_mcount_nc
+ */
+
+.macro mcount_enter
+	stmdb	sp!, {r0-r3, lr}
+.endm
+
+.macro mcount_get_lr reg
+	ldr	\reg, [sp, #20]
+.endm
+
+.macro mcount_exit
+	ldmia	sp!, {r0-r3, ip, lr}
+	mov	pc, ip
+.endm
+
+ENTRY(__gnu_mcount_nc)
+#ifdef CONFIG_DYNAMIC_FTRACE
+	mov	ip, lr
+	ldmia	sp!, {lr}
+	mov	pc, ip
+#else
+	__mcount
+#endif
+ENDPROC(__gnu_mcount_nc)
+
+#ifdef CONFIG_DYNAMIC_FTRACE
+ENTRY(ftrace_caller)
+	__ftrace_caller
+ENDPROC(ftrace_caller)
+#endif
+
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+ENTRY(ftrace_graph_caller)
+	__ftrace_graph_caller
+ENDPROC(ftrace_graph_caller)
+#endif
+
+.purgem mcount_enter
+.purgem mcount_get_lr
+.purgem mcount_exit
+
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+	.globl return_to_handler
+return_to_handler:
+	stmdb	sp!, {r0-r3}
+	mov	r0, fp			@ frame pointer
+	bl	ftrace_return_to_handler
+	mov	lr, r0			@ r0 has real ret addr
+	ldmia	sp!, {r0-r3}
+	mov	pc, lr
+#endif
 
 ENTRY(ftrace_stub)
 .Lftrace_stub:
diff --git a/arch/arm/kernel/ftrace.c b/arch/arm/kernel/ftrace.c
index 971ac8c36ea7..c0062ad1e847 100644
--- a/arch/arm/kernel/ftrace.c
+++ b/arch/arm/kernel/ftrace.c
@@ -24,6 +24,7 @@
 #define	NOP		0xe8bd4000	/* pop {lr} */
 #endif
 
+#ifdef CONFIG_DYNAMIC_FTRACE
 #ifdef CONFIG_OLD_MCOUNT
 #define OLD_MCOUNT_ADDR	((unsigned long) mcount)
 #define OLD_FTRACE_ADDR ((unsigned long) ftrace_caller_old)
@@ -59,9 +60,9 @@ static unsigned long adjust_address(struct dyn_ftrace *rec, unsigned long addr)
 }
 #endif
 
-/* construct a branch (BL) instruction to addr */
 #ifdef CONFIG_THUMB2_KERNEL
-static unsigned long ftrace_call_replace(unsigned long pc, unsigned long addr)
+static unsigned long ftrace_gen_branch(unsigned long pc, unsigned long addr,
+				       bool link)
 {
 	unsigned long s, j1, j2, i1, i2, imm10, imm11;
 	unsigned long first, second;
@@ -83,15 +84,22 @@ static unsigned long ftrace_call_replace(unsigned long pc, unsigned long addr)
 	j2 = (!i2) ^ s;
 
 	first = 0xf000 | (s << 10) | imm10;
-	second = 0xd000 | (j1 << 13) | (j2 << 11) | imm11;
+	second = 0x9000 | (j1 << 13) | (j2 << 11) | imm11;
+	if (link)
+		second |= 1 << 14;
 
 	return (second << 16) | first;
 }
 #else
-static unsigned long ftrace_call_replace(unsigned long pc, unsigned long addr)
+static unsigned long ftrace_gen_branch(unsigned long pc, unsigned long addr,
+				       bool link)
 {
+	unsigned long opcode = 0xea000000;
 	long offset;
 
+	if (link)
+		opcode |= 1 << 24;
+
 	offset = (long)addr - (long)(pc + 8);
 	if (unlikely(offset < -33554432 || offset > 33554428)) {
 		/* Can't generate branches that far (from ARM ARM). Ftrace
@@ -103,10 +111,15 @@ static unsigned long ftrace_call_replace(unsigned long pc, unsigned long addr)
 
 	offset = (offset >> 2) & 0x00ffffff;
 
-	return 0xeb000000 | offset;
+	return opcode | offset;
 }
 #endif
 
+static unsigned long ftrace_call_replace(unsigned long pc, unsigned long addr)
+{
+	return ftrace_gen_branch(pc, addr, true);
+}
+
 static int ftrace_modify_code(unsigned long pc, unsigned long old,
 			      unsigned long new)
 {
@@ -193,3 +206,83 @@ int __init ftrace_dyn_arch_init(void *data)
 
 	return 0;
 }
+#endif /* CONFIG_DYNAMIC_FTRACE */
+
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr,
+			   unsigned long frame_pointer)
+{
+	unsigned long return_hooker = (unsigned long) &return_to_handler;
+	struct ftrace_graph_ent trace;
+	unsigned long old;
+	int err;
+
+	if (unlikely(atomic_read(&current->tracing_graph_pause)))
+		return;
+
+	old = *parent;
+	*parent = return_hooker;
+
+	err = ftrace_push_return_trace(old, self_addr, &trace.depth,
+				       frame_pointer);
+	if (err == -EBUSY) {
+		*parent = old;
+		return;
+	}
+
+	trace.func = self_addr;
+
+	/* Only trace if the calling function expects to */
+	if (!ftrace_graph_entry(&trace)) {
+		current->curr_ret_stack--;
+		*parent = old;
+	}
+}
+
+#ifdef CONFIG_DYNAMIC_FTRACE
+extern unsigned long ftrace_graph_call;
+extern unsigned long ftrace_graph_call_old;
+extern void ftrace_graph_caller_old(void);
+
+static int __ftrace_modify_caller(unsigned long *callsite,
+				  void (*func) (void), bool enable)
+{
+	unsigned long caller_fn = (unsigned long) func;
+	unsigned long pc = (unsigned long) callsite;
+	unsigned long branch = ftrace_gen_branch(pc, caller_fn, false);
+	unsigned long nop = 0xe1a00000;	/* mov r0, r0 */
+	unsigned long old = enable ? nop : branch;
+	unsigned long new = enable ? branch : nop;
+
+	return ftrace_modify_code(pc, old, new);
+}
+
+static int ftrace_modify_graph_caller(bool enable)
+{
+	int ret;
+
+	ret = __ftrace_modify_caller(&ftrace_graph_call,
+				     ftrace_graph_caller,
+				     enable);
+
+#ifdef CONFIG_OLD_MCOUNT
+	if (!ret)
+		ret = __ftrace_modify_caller(&ftrace_graph_call_old,
+					     ftrace_graph_caller_old,
+					     enable);
+#endif
+
+	return ret;
+}
+
+int ftrace_enable_ftrace_graph_caller(void)
+{
+	return ftrace_modify_graph_caller(true);
+}
+
+int ftrace_disable_ftrace_graph_caller(void)
+{
+	return ftrace_modify_graph_caller(false);
+}
+#endif /* CONFIG_DYNAMIC_FTRACE */
+#endif /* CONFIG_FUNCTION_GRAPH_TRACER */
diff --git a/arch/arm/kernel/irq.c b/arch/arm/kernel/irq.c
index 36ad3be4692a..6d616333340f 100644
--- a/arch/arm/kernel/irq.c
+++ b/arch/arm/kernel/irq.c
@@ -35,6 +35,7 @@
 #include <linux/list.h>
 #include <linux/kallsyms.h>
 #include <linux/proc_fs.h>
+#include <linux/ftrace.h>
 
 #include <asm/system.h>
 #include <asm/mach/irq.h>
@@ -105,7 +106,8 @@ unlock:
  * come via this function.  Instead, they should provide their
  * own 'handler'
  */
-asmlinkage void __exception asm_do_IRQ(unsigned int irq, struct pt_regs *regs)
+asmlinkage void __exception_irq_entry
+asm_do_IRQ(unsigned int irq, struct pt_regs *regs)
 {
 	struct pt_regs *old_regs = set_irq_regs(regs);
 
diff --git a/arch/arm/kernel/smp.c b/arch/arm/kernel/smp.c
index 8c1959590252..bbca89872c18 100644
--- a/arch/arm/kernel/smp.c
+++ b/arch/arm/kernel/smp.c
@@ -16,6 +16,7 @@
 #include <linux/cache.h>
 #include <linux/profile.h>
 #include <linux/errno.h>
+#include <linux/ftrace.h>
 #include <linux/mm.h>
 #include <linux/err.h>
 #include <linux/cpu.h>
@@ -457,7 +458,7 @@ static void ipi_timer(void)
 }
 
 #ifdef CONFIG_LOCAL_TIMERS
-asmlinkage void __exception do_local_timer(struct pt_regs *regs)
+asmlinkage void __exception_irq_entry do_local_timer(struct pt_regs *regs)
 {
 	struct pt_regs *old_regs = set_irq_regs(regs);
 	int cpu = smp_processor_id();
@@ -544,7 +545,7 @@ static void ipi_cpu_stop(unsigned int cpu)
  *
  *  Bit 0 - Inter-processor function call
  */
-asmlinkage void __exception do_IPI(struct pt_regs *regs)
+asmlinkage void __exception_irq_entry do_IPI(struct pt_regs *regs)
 {
 	unsigned int cpu = smp_processor_id();
 	struct ipi_data *ipi = &per_cpu(ipi_data, cpu);
diff --git a/arch/arm/kernel/vmlinux.lds.S b/arch/arm/kernel/vmlinux.lds.S
index cead8893b46b..897c1a8f1694 100644
--- a/arch/arm/kernel/vmlinux.lds.S
+++ b/arch/arm/kernel/vmlinux.lds.S
@@ -101,6 +101,7 @@ SECTIONS
 			__exception_text_start = .;
 			*(.exception.text)
 			__exception_text_end = .;
+			IRQENTRY_TEXT
 			TEXT_TEXT
 			SCHED_TEXT
 			LOCK_TEXT
diff --git a/arch/arm/plat-versatile/sched-clock.c b/arch/arm/plat-versatile/sched-clock.c
index 9768cf7e83d7..9696ddc238c9 100644
--- a/arch/arm/plat-versatile/sched-clock.c
+++ b/arch/arm/plat-versatile/sched-clock.c
@@ -20,6 +20,7 @@
  */
 #include <linux/cnt32_to_63.h>
 #include <linux/io.h>
+#include <linux/sched.h>
 #include <asm/div64.h>
 
 #include <mach/hardware.h>