6 files changed, 379 insertions, 126 deletions

diff --git a/arch/sh/oprofile/Makefile b/arch/sh/oprofile/Makefile
index 2efc2e79fd2..8e6eec91c14 100644
--- a/arch/sh/oprofile/Makefile
+++ b/arch/sh/oprofile/Makefile
@@ -6,13 +6,8 @@ DRIVER_OBJS = $(addprefix ../../../drivers/oprofile/, \
 		oprofilefs.o oprofile_stats.o \
 		timer_int.o )
 
-profdrvr-y				:= op_model_null.o
+oprofile-y	:= $(DRIVER_OBJS) common.o backtrace.o
 
-# SH7750-style performance counters exist across 7750/7750S and 7091.
-profdrvr-$(CONFIG_CPU_SUBTYPE_SH7750S)	:= op_model_sh7750.o
-profdrvr-$(CONFIG_CPU_SUBTYPE_SH7750)	:= op_model_sh7750.o
-profdrvr-$(CONFIG_CPU_SUBTYPE_SH7091)	:= op_model_sh7750.o
-
-oprofile-y				:= $(DRIVER_OBJS) $(profdrvr-y)
-
-EXTRA_CFLAGS += -Werror
+oprofile-$(CONFIG_CPU_SUBTYPE_SH7750S)	+= op_model_sh7750.o
+oprofile-$(CONFIG_CPU_SUBTYPE_SH7750)	+= op_model_sh7750.o
+oprofile-$(CONFIG_CPU_SUBTYPE_SH7091)	+= op_model_sh7750.o
diff --git a/arch/sh/oprofile/backtrace.c b/arch/sh/oprofile/backtrace.c
new file mode 100644
index 00000000000..9499a2914f8
--- /dev/null
+++ b/arch/sh/oprofile/backtrace.c
@@ -0,0 +1,114 @@
+/*
+ * SH specific backtracing code for oprofile
+ *
+ * Copyright 2007 STMicroelectronics Ltd.
+ *
+ * Author: Dave Peverley <dpeverley@mpc-data.co.uk>
+ *
+ * Based on ARM oprofile backtrace code by Richard Purdie and in turn, 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/kallsyms.h>
+#include <linux/mm.h>
+#include <asm/ptrace.h>
+#include <asm/uaccess.h>
+#include <asm/sections.h>
+
+/* Limit to stop backtracing too far. */
+static int backtrace_limit = 20;
+
+static unsigned long *
+user_backtrace(unsigned long *stackaddr, struct pt_regs *regs)
+{
+	unsigned long buf_stack;
+
+	/* Also check accessibility of address */
+	if (!access_ok(VERIFY_READ, stackaddr, sizeof(unsigned long)))
+		return NULL;
+
+	if (__copy_from_user_inatomic(&buf_stack, stackaddr, sizeof(unsigned long)))
+		return NULL;
+
+	/* Quick paranoia check */
+	if (buf_stack & 3)
+		return NULL;
+
+	oprofile_add_trace(buf_stack);
+
+	stackaddr++;
+
+	return stackaddr;
+}
+
+/*
+ * |             | /\ Higher addresses
+ * |             |
+ * --------------- stack base (address of current_thread_info)
+ * | thread info |
+ * .             .
+ * |    stack    |
+ * --------------- saved regs->regs[15] value if valid
+ * .             .
+ * --------------- struct pt_regs stored on stack (struct pt_regs *)
+ * |             |
+ * .             .
+ * |             |
+ * --------------- ???
+ * |             |
+ * |             | \/ Lower addresses
+ *
+ * Thus, &pt_regs <-> stack base restricts the valid(ish) fp values
+ */
+static int valid_kernel_stack(unsigned long *stackaddr, struct pt_regs *regs)
+{
+	unsigned long stack = (unsigned long)regs;
+	unsigned long stack_base = (stack & ~(THREAD_SIZE - 1)) + THREAD_SIZE;
+
+	return ((unsigned long)stackaddr > stack) && ((unsigned long)stackaddr < stack_base);
+}
+
+static unsigned long *
+kernel_backtrace(unsigned long *stackaddr, struct pt_regs *regs)
+{
+	unsigned long addr;
+
+	/*
+	 * If not a valid kernel address, keep going till we find one
+	 * or the SP stops being a valid address.
+	 */
+	do {
+		addr = *stackaddr++;
+		oprofile_add_trace(addr);
+	} while (valid_kernel_stack(stackaddr, regs));
+
+	return stackaddr;
+}
+
+void sh_backtrace(struct pt_regs * const regs, unsigned int depth)
+{
+	unsigned long *stackaddr;
+
+	/*
+	 * Paranoia - clip max depth as we could get lost in the weeds.
+	 */
+	if (depth > backtrace_limit)
+		depth = backtrace_limit;
+
+	stackaddr = (unsigned long *)regs->regs[15];
+	if (!user_mode(regs)) {
+		while (depth-- && valid_kernel_stack(stackaddr, regs))
+			stackaddr = kernel_backtrace(stackaddr, regs);
+
+		return;
+	}
+
+	while (depth-- && (stackaddr != NULL))
+		stackaddr = user_backtrace(stackaddr, regs);
+}
diff --git a/arch/sh/oprofile/common.c b/arch/sh/oprofile/common.c
new file mode 100644
index 00000000000..1d97d64cb95
--- /dev/null
+++ b/arch/sh/oprofile/common.c
@@ -0,0 +1,150 @@
+/*
+ * arch/sh/oprofile/init.c
+ *
+ * Copyright (C) 2003 - 2008  Paul Mundt
+ *
+ * Based on arch/mips/oprofile/common.c:
+ *
+ *	Copyright (C) 2004, 2005 Ralf Baechle
+ *	Copyright (C) 2005 MIPS Technologies, Inc.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+#include <linux/kernel.h>
+#include <linux/oprofile.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/smp.h>
+#include <asm/processor.h>
+#include "op_impl.h"
+
+extern struct op_sh_model op_model_sh7750_ops __weak;
+extern struct op_sh_model op_model_sh4a_ops __weak;
+
+static struct op_sh_model *model;
+
+static struct op_counter_config ctr[20];
+
+extern void sh_backtrace(struct pt_regs * const regs, unsigned int depth);
+
+static int op_sh_setup(void)
+{
+	/* Pre-compute the values to stuff in the hardware registers.  */
+	model->reg_setup(ctr);
+
+	/* Configure the registers on all cpus.  */
+	on_each_cpu(model->cpu_setup, NULL, 1);
+
+        return 0;
+}
+
+static int op_sh_create_files(struct super_block *sb, struct dentry *root)
+{
+	int i, ret = 0;
+
+	for (i = 0; i < model->num_counters; i++) {
+		struct dentry *dir;
+		char buf[4];
+
+		snprintf(buf, sizeof(buf), "%d", i);
+		dir = oprofilefs_mkdir(sb, root, buf);
+
+		ret |= oprofilefs_create_ulong(sb, dir, "enabled", &ctr[i].enabled);
+		ret |= oprofilefs_create_ulong(sb, dir, "event", &ctr[i].event);
+		ret |= oprofilefs_create_ulong(sb, dir, "kernel", &ctr[i].kernel);
+		ret |= oprofilefs_create_ulong(sb, dir, "user", &ctr[i].user);
+
+		if (model->create_files)
+			ret |= model->create_files(sb, dir);
+		else
+			ret |= oprofilefs_create_ulong(sb, dir, "count", &ctr[i].count);
+
+		/* Dummy entries */
+		ret |= oprofilefs_create_ulong(sb, dir, "unit_mask", &ctr[i].unit_mask);
+	}
+
+	return ret;
+}
+
+static int op_sh_start(void)
+{
+	/* Enable performance monitoring for all counters.  */
+	on_each_cpu(model->cpu_start, NULL, 1);
+
+	return 0;
+}
+
+static void op_sh_stop(void)
+{
+	/* Disable performance monitoring for all counters.  */
+	on_each_cpu(model->cpu_stop, NULL, 1);
+}
+
+int __init oprofile_arch_init(struct oprofile_operations *ops)
+{
+	struct op_sh_model *lmodel = NULL;
+	int ret;
+
+	/*
+	 * Always assign the backtrace op. If the counter initialization
+	 * fails, we fall back to the timer which will still make use of
+	 * this.
+	 */
+	ops->backtrace = sh_backtrace;
+
+	switch (current_cpu_data.type) {
+	/* SH-4 types */
+	case CPU_SH7750:
+	case CPU_SH7750S:
+		lmodel = &op_model_sh7750_ops;
+		break;
+
+        /* SH-4A types */
+	case CPU_SH7763:
+	case CPU_SH7770:
+	case CPU_SH7780:
+	case CPU_SH7781:
+	case CPU_SH7785:
+	case CPU_SH7723:
+	case CPU_SHX3:
+		lmodel = &op_model_sh4a_ops;
+		break;
+
+	/* SH4AL-DSP types */
+	case CPU_SH7343:
+	case CPU_SH7722:
+	case CPU_SH7366:
+		lmodel = &op_model_sh4a_ops;
+		break;
+	}
+
+	if (!lmodel)
+		return -ENODEV;
+	if (!(current_cpu_data.flags & CPU_HAS_PERF_COUNTER))
+		return -ENODEV;
+
+	ret = lmodel->init();
+	if (unlikely(ret != 0))
+		return ret;
+
+	model = lmodel;
+
+	ops->setup		= op_sh_setup;
+	ops->create_files	= op_sh_create_files;
+	ops->start		= op_sh_start;
+	ops->stop		= op_sh_stop;
+	ops->cpu_type		= lmodel->cpu_type;
+
+	printk(KERN_INFO "oprofile: using %s performance monitoring.\n",
+	       lmodel->cpu_type);
+
+	return 0;
+}
+
+void oprofile_arch_exit(void)
+{
+	if (model && model->exit)
+		model->exit();
+}
diff --git a/arch/sh/oprofile/op_impl.h b/arch/sh/oprofile/op_impl.h
new file mode 100644
index 00000000000..4d509975eba
--- /dev/null
+++ b/arch/sh/oprofile/op_impl.h
@@ -0,0 +1,33 @@
+#ifndef __OP_IMPL_H
+#define __OP_IMPL_H
+
+/* Per-counter configuration as set via oprofilefs.  */
+struct op_counter_config {
+	unsigned long enabled;
+	unsigned long event;
+
+	unsigned long long count;
+
+	/* Dummy values for userspace tool compliance */
+	unsigned long kernel;
+	unsigned long user;
+	unsigned long unit_mask;
+};
+
+/* Per-architecture configury and hooks.  */
+struct op_sh_model {
+	void (*reg_setup)(struct op_counter_config *);
+	int (*create_files)(struct super_block *sb, struct dentry *dir);
+	void (*cpu_setup)(void *dummy);
+	int (*init)(void);
+	void (*exit)(void);
+	void (*cpu_start)(void *args);
+	void (*cpu_stop)(void *args);
+	char *cpu_type;
+	unsigned char num_counters;
+};
+
+/* arch/sh/oprofile/common.c */
+extern void sh_backtrace(struct pt_regs * const regs, unsigned int depth);
+
+#endif /* __OP_IMPL_H */
diff --git a/arch/sh/oprofile/op_model_null.c b/arch/sh/oprofile/op_model_null.c
deleted file mode 100644
index a845b088edb..00000000000
--- a/arch/sh/oprofile/op_model_null.c
+++ /dev/null
@@ -1,23 +0,0 @@
-/*
- * arch/sh/oprofile/op_model_null.c
- *
- * Copyright (C) 2003  Paul Mundt
- *
- * This file is subject to the terms and conditions of the GNU General Public
- * License.  See the file "COPYING" in the main directory of this archive
- * for more details.
- */
-#include <linux/kernel.h>
-#include <linux/oprofile.h>
-#include <linux/init.h>
-#include <linux/errno.h>
-
-int __init oprofile_arch_init(struct oprofile_operations *ops)
-{
-	return -ENODEV;
-}
-
-void oprofile_arch_exit(void)
-{
-}
-
diff --git a/arch/sh/oprofile/op_model_sh7750.c b/arch/sh/oprofile/op_model_sh7750.c
index 008b3b03750..c892c7c30c2 100644
--- a/arch/sh/oprofile/op_model_sh7750.c
+++ b/arch/sh/oprofile/op_model_sh7750.c
@@ -3,7 +3,7 @@
  *
  * OProfile support for SH7750/SH7750S Performance Counters
  *
- * Copyright (C) 2003, 2004  Paul Mundt
+ * Copyright (C) 2003 - 2008  Paul Mundt
  *
  * This file is subject to the terms and conditions of the GNU General Public
  * License.  See the file "COPYING" in the main directory of this archive
@@ -15,19 +15,16 @@
 #include <linux/init.h>
 #include <linux/errno.h>
 #include <linux/interrupt.h>
+#include <linux/io.h>
 #include <linux/fs.h>
-#include <asm/uaccess.h>
-#include <asm/io.h>
+#include "op_impl.h"
 
 #define PM_CR_BASE	0xff000084	/* 16-bit */
 #define PM_CTR_BASE	0xff100004	/* 32-bit */
 
-#define PMCR1		(PM_CR_BASE  + 0x00)
-#define PMCR2		(PM_CR_BASE  + 0x04)
-#define PMCTR1H		(PM_CTR_BASE + 0x00)
-#define PMCTR1L		(PM_CTR_BASE + 0x04)
-#define PMCTR2H		(PM_CTR_BASE + 0x08)
-#define PMCTR2L		(PM_CTR_BASE + 0x0c)
+#define PMCR(n)		(PM_CR_BASE + ((n) * 0x04))
+#define PMCTRH(n)	(PM_CTR_BASE + 0x00 + ((n) * 0x08))
+#define PMCTRL(n)	(PM_CTR_BASE + 0x04 + ((n) * 0x08))
 
 #define PMCR_PMM_MASK	0x0000003f
 
@@ -36,25 +33,15 @@
 #define PMCR_PMST	0x00004000
 #define PMCR_PMEN	0x00008000
 
-#define PMCR_ENABLE	(PMCR_PMST | PMCR_PMEN)
+struct op_sh_model op_model_sh7750_ops;
 
-/*
- * SH7750/SH7750S have 2 perf counters
- */
 #define NR_CNTRS	2
 
-struct op_counter_config {
-	unsigned long enabled;
-	unsigned long event;
-	unsigned long count;
-
-	/* Dummy values for userspace tool compliance */
-	unsigned long kernel;
-	unsigned long user;
-	unsigned long unit_mask;
-};
-
-static struct op_counter_config ctr[NR_CNTRS];
+static struct sh7750_ppc_register_config {
+	unsigned int ctrl;
+	unsigned long cnt_hi;
+	unsigned long cnt_lo;
+} regcache[NR_CNTRS];
 
 /*
  * There are a number of events supported by each counter (33 in total).
@@ -116,12 +103,8 @@ static int sh7750_timer_notify(struct pt_regs *regs)
 
 static u64 sh7750_read_counter(int counter)
 {
-	u32 hi, lo;
-
-	hi = (counter == 0) ? ctrl_inl(PMCTR1H) : ctrl_inl(PMCTR2H);
-	lo = (counter == 0) ? ctrl_inl(PMCTR1L) : ctrl_inl(PMCTR2L);
-
-	return (u64)((u64)(hi & 0xffff) << 32) | lo;
+	return (u64)((u64)(__raw_readl(PMCTRH(counter)) & 0xffff) << 32) |
+			   __raw_readl(PMCTRL(counter));
 }
 
 /*
@@ -170,11 +153,7 @@ static ssize_t sh7750_write_count(struct file *file, const char __user *buf,
 	 */
 	WARN_ON(val != 0);
 
-	if (counter == 0) {
-		ctrl_outw(ctrl_inw(PMCR1) | PMCR_PMCLR, PMCR1);
-	} else {
-		ctrl_outw(ctrl_inw(PMCR2) | PMCR_PMCLR, PMCR2);
-	}
+	__raw_writew(__raw_readw(PMCR(counter)) | PMCR_PMCLR, PMCR(counter));
 
 	return count;
 }
@@ -184,88 +163,93 @@ static const struct file_operations count_fops = {
 	.write		= sh7750_write_count,
 };
 
-static int sh7750_perf_counter_create_files(struct super_block *sb, struct dentry *root)
+static int sh7750_ppc_create_files(struct super_block *sb, struct dentry *dir)
 {
-	int i;
+	return oprofilefs_create_file(sb, dir, "count", &count_fops);
+}
 
-	for (i = 0; i < NR_CNTRS; i++) {
-		struct dentry *dir;
-		char buf[4];
+static void sh7750_ppc_reg_setup(struct op_counter_config *ctr)
+{
+	unsigned int counters = op_model_sh7750_ops.num_counters;
+	int i;
 
-		snprintf(buf, sizeof(buf), "%d", i);
-		dir = oprofilefs_mkdir(sb, root, buf);
+	for (i = 0; i < counters; i++) {
+		regcache[i].ctrl	= 0;
+		regcache[i].cnt_hi	= 0;
+		regcache[i].cnt_lo	= 0;
 
-		oprofilefs_create_ulong(sb, dir, "enabled", &ctr[i].enabled);
-		oprofilefs_create_ulong(sb, dir, "event", &ctr[i].event);
-		oprofilefs_create_file(sb, dir, "count", &count_fops);
+		if (!ctr[i].enabled)
+			continue;
 
-		/* Dummy entries */
-		oprofilefs_create_ulong(sb, dir, "kernel", &ctr[i].kernel);
-		oprofilefs_create_ulong(sb, dir, "user", &ctr[i].user);
-		oprofilefs_create_ulong(sb, dir, "unit_mask", &ctr[i].unit_mask);
+		regcache[i].ctrl |= ctr[i].event | PMCR_PMEN | PMCR_PMST;
+		regcache[i].cnt_hi = (unsigned long)((ctr->count >> 32) & 0xffff);
+		regcache[i].cnt_lo = (unsigned long)(ctr->count & 0xffffffff);
 	}
-
-	return 0;
 }
 
-static int sh7750_perf_counter_start(void)
+static void sh7750_ppc_cpu_setup(void *args)
 {
-	u16 pmcr;
-
-	/* Enable counter 1 */
-	if (ctr[0].enabled) {
-		pmcr = ctrl_inw(PMCR1);
-		WARN_ON(pmcr & PMCR_PMEN);
-
-		pmcr &= ~PMCR_PMM_MASK;
-		pmcr |= ctr[0].event;
-		ctrl_outw(pmcr | PMCR_ENABLE, PMCR1);
-	}
-
-	/* Enable counter 2 */
-	if (ctr[1].enabled) {
-		pmcr = ctrl_inw(PMCR2);
-		WARN_ON(pmcr & PMCR_PMEN);
+	unsigned int counters = op_model_sh7750_ops.num_counters;
+	int i;
 
-		pmcr &= ~PMCR_PMM_MASK;
-		pmcr |= ctr[1].event;
-		ctrl_outw(pmcr | PMCR_ENABLE, PMCR2);
+	for (i = 0; i < counters; i++) {
+		__raw_writew(0, PMCR(i));
+		__raw_writel(regcache[i].cnt_hi, PMCTRH(i));
+		__raw_writel(regcache[i].cnt_lo, PMCTRL(i));
 	}
-
-	return register_timer_hook(sh7750_timer_notify);
 }
 
-static void sh7750_perf_counter_stop(void)
+static void sh7750_ppc_cpu_start(void *args)
 {
-	ctrl_outw(ctrl_inw(PMCR1) & ~PMCR_PMEN, PMCR1);
-	ctrl_outw(ctrl_inw(PMCR2) & ~PMCR_PMEN, PMCR2);
+	unsigned int counters = op_model_sh7750_ops.num_counters;
+	int i;
 
-	unregister_timer_hook(sh7750_timer_notify);
+	for (i = 0; i < counters; i++)
+		__raw_writew(regcache[i].ctrl, PMCR(i));
 }
 
-static struct oprofile_operations sh7750_perf_counter_ops = {
-	.create_files	= sh7750_perf_counter_create_files,
-	.start		= sh7750_perf_counter_start,
-	.stop		= sh7750_perf_counter_stop,
-};
-
-int __init oprofile_arch_init(struct oprofile_operations *ops)
+static void sh7750_ppc_cpu_stop(void *args)
 {
-	if (!(current_cpu_data.flags & CPU_HAS_PERF_COUNTER))
-		return -ENODEV;
+	unsigned int counters = op_model_sh7750_ops.num_counters;
+	int i;
 
-	ops = &sh7750_perf_counter_ops;
-	ops->cpu_type = "sh/sh7750";
+	/* Disable the counters */
+	for (i = 0; i < counters; i++)
+		__raw_writew(__raw_readw(PMCR(i)) & ~PMCR_PMEN, PMCR(i));
+}
 
-	printk(KERN_INFO "oprofile: using SH-4 performance monitoring.\n");
+static inline void sh7750_ppc_reset(void)
+{
+	unsigned int counters = op_model_sh7750_ops.num_counters;
+	int i;
 
 	/* Clear the counters */
-	ctrl_outw(ctrl_inw(PMCR1) | PMCR_PMCLR, PMCR1);
-	ctrl_outw(ctrl_inw(PMCR2) | PMCR_PMCLR, PMCR2);
+	for (i = 0; i < counters; i++)
+		__raw_writew(__raw_readw(PMCR(i)) | PMCR_PMCLR, PMCR(i));
+}
 
-	return 0;
+static int sh7750_ppc_init(void)
+{
+	sh7750_ppc_reset();
+
+	return register_timer_hook(sh7750_timer_notify);
 }
 
-void oprofile_arch_exit(void)
+static void sh7750_ppc_exit(void)
 {
+	unregister_timer_hook(sh7750_timer_notify);
+
+	sh7750_ppc_reset();
 }
+
+struct op_sh_model op_model_sh7750_ops = {
+	.cpu_type	= "sh/sh7750",
+	.num_counters	= NR_CNTRS,
+	.reg_setup	= sh7750_ppc_reg_setup,
+	.cpu_setup	= sh7750_ppc_cpu_setup,
+	.cpu_start	= sh7750_ppc_cpu_start,
+	.cpu_stop	= sh7750_ppc_cpu_stop,
+	.init		= sh7750_ppc_init,
+	.exit		= sh7750_ppc_exit,
+	.create_files	= sh7750_ppc_create_files,
+};