[PATCH] ARM: 2761/1: OProfile: Add call graphing support for arm

Patch from Richard Purdie

Add functions to generate backtraces of both kernel and user processes
which allows oprofile's call graphing functionality to be used on arm.
This requires unstripped binaries/libs which use a frame pointer.

Signed-off-by: Richard Purdie
Signed-off-by: Zwane Mwaikambo <zwane@arm.linux.org.uk>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

4 files changed, 149 insertions, 1 deletions

diff --git a/arch/arm/oprofile/Makefile b/arch/arm/oprofile/Makefile
index ba1a6e9..8ffb523 100644
--- a/arch/arm/oprofile/Makefile
+++ b/arch/arm/oprofile/Makefile
@@ -6,6 +6,6 @@ DRIVER_OBJS = $(addprefix ../../../drivers/oprofile/, \
 		oprofilefs.o oprofile_stats.o \
 		timer_int.o )
 
-oprofile-y				:= $(DRIVER_OBJS) init.o
+oprofile-y				:= $(DRIVER_OBJS) init.o backtrace.o
 oprofile-$(CONFIG_CPU_XSCALE)		+= common.o op_model_xscale.o
 
diff --git a/arch/arm/oprofile/backtrace.c b/arch/arm/oprofile/backtrace.c
new file mode 100644
index 0000000..ec58d3e
--- /dev/null
+++ b/arch/arm/oprofile/backtrace.c
@@ -0,0 +1,144 @@
+/*
+ * Arm specific backtracing code for oprofile
+ *
+ * Copyright 2005 Openedhand Ltd.
+ *
+ * Author: Richard Purdie <rpurdie@openedhand.com>
+ *
+ * Based on i386 oprofile backtrace code by John Levon, David Smith
+ *
+ * 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
+ * published by the Free Software Foundation.
+ *
+ */
+
+#include <linux/oprofile.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <asm/ptrace.h>
+#include <asm/uaccess.h>
+
+
+/*
+ * The registers we're interested in are at the end of the variable
+ * length saved register structure. The fp points at the end of this
+ * structure so the address of this struct is:
+ * (struct frame_tail *)(xxx->fp)-1
+ */
+struct frame_tail {
+	struct frame_tail *fp;
+	unsigned long sp;
+	unsigned long lr;
+} __attribute__((packed));
+
+
+#ifdef CONFIG_FRAME_POINTER
+static struct frame_tail* kernel_backtrace(struct frame_tail *tail)
+{
+	oprofile_add_trace(tail->lr);
+
+	/* frame pointers should strictly progress back up the stack
+	 * (towards higher addresses) */
+	if (tail >= tail->fp)
+		return NULL;
+
+	return tail->fp-1;
+}
+#endif
+
+static struct frame_tail* user_backtrace(struct frame_tail *tail)
+{
+	struct frame_tail buftail;
+
+	/* hardware pte might not be valid due to dirty/accessed bit emulation
+	 * so we use copy_from_user and benefit from exception fixups */
+	if (copy_from_user(&buftail, tail, sizeof(struct frame_tail)))
+		return NULL;
+
+	oprofile_add_trace(buftail.lr);
+
+	/* frame pointers should strictly progress back up the stack
+	 * (towards higher addresses) */
+	if (tail >= buftail.fp)
+		return NULL;
+
+	return buftail.fp-1;
+}
+
+/* Compare two addresses and see if they're on the same page */
+#define CMP_ADDR_EQUAL(x,y,offset) ((((unsigned long) x) >> PAGE_SHIFT) \
+	== ((((unsigned long) y) + offset) >> PAGE_SHIFT))
+
+/* check that the page(s) containing the frame tail are present */
+static int pages_present(struct frame_tail *tail)
+{
+	struct mm_struct * mm = current->mm;
+
+	if (!check_user_page_readable(mm, (unsigned long)tail))
+		return 0;
+
+	if (CMP_ADDR_EQUAL(tail, tail, 8))
+		return 1;
+
+	if (!check_user_page_readable(mm, ((unsigned long)tail) + 8))
+		return 0;
+
+	return 1;
+}
+
+/*
+ * |             | /\ Higher addresses
+ * |             |
+ * --------------- stack base (address of current_thread_info)
+ * | thread info |
+ * .             .
+ * |    stack    |
+ * --------------- saved regs->ARM_fp value if valid (frame_tail address)
+ * .             .
+ * --------------- struct pt_regs stored on stack (struct pt_regs *)
+ * |             |
+ * .             .
+ * |             |
+ * --------------- %esp
+ * |             |
+ * |             | \/ Lower addresses
+ *
+ * Thus, &pt_regs <-> stack base restricts the valid(ish) fp values
+ */
+static int valid_kernel_stack(struct frame_tail *tail, struct pt_regs *regs)
+{
+	unsigned long tailaddr = (unsigned long)tail;
+	unsigned long stack = (unsigned long)regs;
+	unsigned long stack_base = (stack & ~(THREAD_SIZE - 1)) + THREAD_SIZE;
+
+	return (tailaddr > stack) && (tailaddr < stack_base);
+}
+
+void arm_backtrace(struct pt_regs const *regs, unsigned int depth)
+{
+	struct frame_tail *tail;
+	unsigned long last_address = 0;
+
+	tail = ((struct frame_tail *) regs->ARM_fp) - 1;
+
+	if (!user_mode(regs)) {
+
+#ifdef CONFIG_FRAME_POINTER
+		while (depth-- && tail && valid_kernel_stack(tail, regs)) {
+			tail = kernel_backtrace(tail);
+		}
+#endif
+		return;
+	}
+
+	while (depth-- && tail && !((unsigned long) tail & 3)) {
+		if ((!CMP_ADDR_EQUAL(last_address, tail, 0)
+			|| !CMP_ADDR_EQUAL(last_address, tail, 8))
+				&& !pages_present(tail))
+			return;
+		last_address = (unsigned long) tail;
+		tail = user_backtrace(tail);
+	}
+}
+
diff --git a/arch/arm/oprofile/init.c b/arch/arm/oprofile/init.c
index cce3d30..d315a3a 100644
--- a/arch/arm/oprofile/init.c
+++ b/arch/arm/oprofile/init.c
@@ -20,6 +20,8 @@ int __init oprofile_arch_init(struct oprofile_operations *ops)
 	ret = pmu_init(ops, &op_xscale_spec);
 #endif
 
+	ops->backtrace = arm_backtrace;
+
 	return ret;
 }
 
diff --git a/arch/arm/oprofile/op_arm_model.h b/arch/arm/oprofile/op_arm_model.h
index 2d4caf4..2148d07 100644
--- a/arch/arm/oprofile/op_arm_model.h
+++ b/arch/arm/oprofile/op_arm_model.h
@@ -24,6 +24,8 @@ struct op_arm_model_spec {
 extern struct op_arm_model_spec op_xscale_spec;
 #endif
 
+extern void arm_backtrace(struct pt_regs * const regs, unsigned int depth);
+
 extern int __init pmu_init(struct oprofile_operations *ops, struct op_arm_model_spec *spec);
 extern void pmu_exit(void);
 #endif /* OP_ARM_MODEL_H */