Merge branch 'oprofile/ring_buffer' into oprofile/oprofile-for-tip

13 files changed, 566 insertions, 465 deletions

diff --git a/arch/Kconfig b/arch/Kconfig
index 471e72d..2e13aa2 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -6,6 +6,8 @@ config OPROFILE
 	tristate "OProfile system profiling (EXPERIMENTAL)"
 	depends on PROFILING
 	depends on HAVE_OPROFILE
+	select TRACING
+	select RING_BUFFER
 	help
 	  OProfile is a profiling system capable of profiling the
 	  whole system, include the kernel, kernel modules, libraries,
diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c
index 5095137..8fdf06e 100644
--- a/arch/x86/oprofile/op_model_amd.c
+++ b/arch/x86/oprofile/op_model_amd.c
@@ -2,7 +2,7 @@
  * @file op_model_amd.c
  * athlon / K7 / K8 / Family 10h model-specific MSR operations
  *
- * @remark Copyright 2002-2008 OProfile authors
+ * @remark Copyright 2002-2009 OProfile authors
  * @remark Read the file COPYING
  *
  * @author John Levon
@@ -10,7 +10,7 @@
  * @author Graydon Hoare
  * @author Robert Richter <robert.richter@amd.com>
  * @author Barry Kasindorf
-*/
+ */
 
 #include <linux/oprofile.h>
 #include <linux/device.h>
@@ -60,56 +60,10 @@ static unsigned long reset_value[NUM_COUNTERS];
 #define IBS_OP_LOW_VALID_BIT		(1ULL<<18)	/* bit 18 */
 #define IBS_OP_LOW_ENABLE		(1ULL<<17)	/* bit 17 */
 
-/* Codes used in cpu_buffer.c */
-/* This produces duplicate code, need to be fixed */
-#define IBS_FETCH_BEGIN 3
-#define IBS_OP_BEGIN    4
-
-/* The function interface needs to be fixed, something like add
-   data. Should then be added to linux/oprofile.h. */
-extern void
-oprofile_add_ibs_sample(struct pt_regs *const regs,
-			unsigned int *const ibs_sample, int ibs_code);
-
-struct ibs_fetch_sample {
-	/* MSRC001_1031 IBS Fetch Linear Address Register */
-	unsigned int ibs_fetch_lin_addr_low;
-	unsigned int ibs_fetch_lin_addr_high;
-	/* MSRC001_1030 IBS Fetch Control Register */
-	unsigned int ibs_fetch_ctl_low;
-	unsigned int ibs_fetch_ctl_high;
-	/* MSRC001_1032 IBS Fetch Physical Address Register */
-	unsigned int ibs_fetch_phys_addr_low;
-	unsigned int ibs_fetch_phys_addr_high;
-};
-
-struct ibs_op_sample {
-	/* MSRC001_1034 IBS Op Logical Address Register (IbsRIP) */
-	unsigned int ibs_op_rip_low;
-	unsigned int ibs_op_rip_high;
-	/* MSRC001_1035 IBS Op Data Register */
-	unsigned int ibs_op_data1_low;
-	unsigned int ibs_op_data1_high;
-	/* MSRC001_1036 IBS Op Data 2 Register */
-	unsigned int ibs_op_data2_low;
-	unsigned int ibs_op_data2_high;
-	/* MSRC001_1037 IBS Op Data 3 Register */
-	unsigned int ibs_op_data3_low;
-	unsigned int ibs_op_data3_high;
-	/* MSRC001_1038 IBS DC Linear Address Register (IbsDcLinAd) */
-	unsigned int ibs_dc_linear_low;
-	unsigned int ibs_dc_linear_high;
-	/* MSRC001_1039 IBS DC Physical Address Register (IbsDcPhysAd) */
-	unsigned int ibs_dc_phys_low;
-	unsigned int ibs_dc_phys_high;
-};
-
-/*
- * unitialize the APIC for the IBS interrupts if needed on AMD Family10h+
-*/
-static void clear_ibs_nmi(void);
+#define IBS_FETCH_SIZE	6
+#define IBS_OP_SIZE	12
 
-static int ibs_allowed;	/* AMD Family10h and later */
+static int has_ibs;	/* AMD Family10h and later */
 
 struct op_ibs_config {
 	unsigned long op_enabled;
@@ -200,31 +154,29 @@ static inline int
 op_amd_handle_ibs(struct pt_regs * const regs,
 		  struct op_msrs const * const msrs)
 {
-	unsigned int low, high;
-	struct ibs_fetch_sample ibs_fetch;
-	struct ibs_op_sample ibs_op;
+	u32 low, high;
+	u64 msr;
+	struct op_entry entry;
 
-	if (!ibs_allowed)
+	if (!has_ibs)
 		return 1;
 
 	if (ibs_config.fetch_enabled) {
 		rdmsr(MSR_AMD64_IBSFETCHCTL, low, high);
 		if (high & IBS_FETCH_HIGH_VALID_BIT) {
-			ibs_fetch.ibs_fetch_ctl_high = high;
-			ibs_fetch.ibs_fetch_ctl_low = low;
-			rdmsr(MSR_AMD64_IBSFETCHLINAD, low, high);
-			ibs_fetch.ibs_fetch_lin_addr_high = high;
-			ibs_fetch.ibs_fetch_lin_addr_low = low;
-			rdmsr(MSR_AMD64_IBSFETCHPHYSAD, low, high);
-			ibs_fetch.ibs_fetch_phys_addr_high = high;
-			ibs_fetch.ibs_fetch_phys_addr_low = low;
-
-			oprofile_add_ibs_sample(regs,
-						(unsigned int *)&ibs_fetch,
-						IBS_FETCH_BEGIN);
-
-			/*reenable the IRQ */
-			rdmsr(MSR_AMD64_IBSFETCHCTL, low, high);
+			rdmsrl(MSR_AMD64_IBSFETCHLINAD, msr);
+			oprofile_write_reserve(&entry, regs, msr,
+					       IBS_FETCH_CODE, IBS_FETCH_SIZE);
+			oprofile_add_data(&entry, (u32)msr);
+			oprofile_add_data(&entry, (u32)(msr >> 32));
+			oprofile_add_data(&entry, low);
+			oprofile_add_data(&entry, high);
+			rdmsrl(MSR_AMD64_IBSFETCHPHYSAD, msr);
+			oprofile_add_data(&entry, (u32)msr);
+			oprofile_add_data(&entry, (u32)(msr >> 32));
+			oprofile_write_commit(&entry);
+
+			/* reenable the IRQ */
 			high &= ~IBS_FETCH_HIGH_VALID_BIT;
 			high |= IBS_FETCH_HIGH_ENABLE;
 			low &= IBS_FETCH_LOW_MAX_CNT_MASK;
@@ -235,30 +187,29 @@ op_amd_handle_ibs(struct pt_regs * const regs,
 	if (ibs_config.op_enabled) {
 		rdmsr(MSR_AMD64_IBSOPCTL, low, high);
 		if (low & IBS_OP_LOW_VALID_BIT) {
-			rdmsr(MSR_AMD64_IBSOPRIP, low, high);
-			ibs_op.ibs_op_rip_low = low;
-			ibs_op.ibs_op_rip_high = high;
-			rdmsr(MSR_AMD64_IBSOPDATA, low, high);
-			ibs_op.ibs_op_data1_low = low;
-			ibs_op.ibs_op_data1_high = high;
-			rdmsr(MSR_AMD64_IBSOPDATA2, low, high);
-			ibs_op.ibs_op_data2_low = low;
-			ibs_op.ibs_op_data2_high = high;
-			rdmsr(MSR_AMD64_IBSOPDATA3, low, high);
-			ibs_op.ibs_op_data3_low = low;
-			ibs_op.ibs_op_data3_high = high;
-			rdmsr(MSR_AMD64_IBSDCLINAD, low, high);
-			ibs_op.ibs_dc_linear_low = low;
-			ibs_op.ibs_dc_linear_high = high;
-			rdmsr(MSR_AMD64_IBSDCPHYSAD, low, high);
-			ibs_op.ibs_dc_phys_low = low;
-			ibs_op.ibs_dc_phys_high = high;
+			rdmsrl(MSR_AMD64_IBSOPRIP, msr);
+			oprofile_write_reserve(&entry, regs, msr,
+					       IBS_OP_CODE, IBS_OP_SIZE);
+			oprofile_add_data(&entry, (u32)msr);
+			oprofile_add_data(&entry, (u32)(msr >> 32));
+			rdmsrl(MSR_AMD64_IBSOPDATA, msr);
+			oprofile_add_data(&entry, (u32)msr);
+			oprofile_add_data(&entry, (u32)(msr >> 32));
+			rdmsrl(MSR_AMD64_IBSOPDATA2, msr);
+			oprofile_add_data(&entry, (u32)msr);
+			oprofile_add_data(&entry, (u32)(msr >> 32));
+			rdmsrl(MSR_AMD64_IBSOPDATA3, msr);
+			oprofile_add_data(&entry, (u32)msr);
+			oprofile_add_data(&entry, (u32)(msr >> 32));
+			rdmsrl(MSR_AMD64_IBSDCLINAD, msr);
+			oprofile_add_data(&entry, (u32)msr);
+			oprofile_add_data(&entry, (u32)(msr >> 32));
+			rdmsrl(MSR_AMD64_IBSDCPHYSAD, msr);
+			oprofile_add_data(&entry, (u32)msr);
+			oprofile_add_data(&entry, (u32)(msr >> 32));
+			oprofile_write_commit(&entry);
 
 			/* reenable the IRQ */
-			oprofile_add_ibs_sample(regs,
-						(unsigned int *)&ibs_op,
-						IBS_OP_BEGIN);
-			rdmsr(MSR_AMD64_IBSOPCTL, low, high);
 			high = 0;
 			low &= ~IBS_OP_LOW_VALID_BIT;
 			low |= IBS_OP_LOW_ENABLE;
@@ -308,14 +259,14 @@ static void op_amd_start(struct op_msrs const * const msrs)
 	}
 
 #ifdef CONFIG_OPROFILE_IBS
-	if (ibs_allowed && ibs_config.fetch_enabled) {
+	if (has_ibs && ibs_config.fetch_enabled) {
 		low = (ibs_config.max_cnt_fetch >> 4) & 0xFFFF;
 		high = ((ibs_config.rand_en & 0x1) << 25) /* bit 57 */
 			+ IBS_FETCH_HIGH_ENABLE;
 		wrmsr(MSR_AMD64_IBSFETCHCTL, low, high);
 	}
 
-	if (ibs_allowed && ibs_config.op_enabled) {
+	if (has_ibs && ibs_config.op_enabled) {
 		low = ((ibs_config.max_cnt_op >> 4) & 0xFFFF)
 			+ ((ibs_config.dispatched_ops & 0x1) << 19) /* bit 19 */
 			+ IBS_OP_LOW_ENABLE;
@@ -331,8 +282,10 @@ static void op_amd_stop(struct op_msrs const * const msrs)
 	unsigned int low, high;
 	int i;
 
-	/* Subtle: stop on all counters to avoid race with
-	 * setting our pm callback */
+	/*
+	 * Subtle: stop on all counters to avoid race with setting our
+	 * pm callback
+	 */
 	for (i = 0 ; i < NUM_COUNTERS ; ++i) {
 		if (!reset_value[i])
 			continue;
@@ -342,14 +295,16 @@ static void op_amd_stop(struct op_msrs const * const msrs)
 	}
 
 #ifdef CONFIG_OPROFILE_IBS
-	if (ibs_allowed && ibs_config.fetch_enabled) {
-		low = 0;		/* clear max count and enable */
+	if (has_ibs && ibs_config.fetch_enabled) {
+		/* clear max count and enable */
+		low = 0;
 		high = 0;
 		wrmsr(MSR_AMD64_IBSFETCHCTL, low, high);
 	}
 
-	if (ibs_allowed && ibs_config.op_enabled) {
-		low = 0;		/* clear max count and enable */
+	if (has_ibs && ibs_config.op_enabled) {
+		/* clear max count and enable */
+		low = 0;
 		high = 0;
 		wrmsr(MSR_AMD64_IBSOPCTL, low, high);
 	}
@@ -370,18 +325,7 @@ static void op_amd_shutdown(struct op_msrs const * const msrs)
 	}
 }
 
-#ifndef CONFIG_OPROFILE_IBS
-
-/* no IBS support */
-
-static int op_amd_init(struct oprofile_operations *ops)
-{
-	return 0;
-}
-
-static void op_amd_exit(void) {}
-
-#else
+#ifdef CONFIG_OPROFILE_IBS
 
 static u8 ibs_eilvt_off;
 
@@ -395,7 +339,7 @@ static inline void apic_clear_ibs_nmi_per_cpu(void *arg)
 	setup_APIC_eilvt_ibs(0, APIC_EILVT_MSG_FIX, 1);
 }
 
-static int pfm_amd64_setup_eilvt(void)
+static int init_ibs_nmi(void)
 {
 #define IBSCTL_LVTOFFSETVAL		(1 << 8)
 #define IBSCTL				0x1cc
@@ -419,6 +363,7 @@ static int pfm_amd64_setup_eilvt(void)
 				       | IBSCTL_LVTOFFSETVAL);
 		pci_read_config_dword(cpu_cfg, IBSCTL, &value);
 		if (value != (ibs_eilvt_off | IBSCTL_LVTOFFSETVAL)) {
+			pci_dev_put(cpu_cfg);
 			printk(KERN_DEBUG "Failed to setup IBS LVT offset, "
 				"IBSCTL = 0x%08x", value);
 			return 1;
@@ -443,33 +388,35 @@ static int pfm_amd64_setup_eilvt(void)
 	return 0;
 }
 
-/*
- * initialize the APIC for the IBS interrupts
- * if available (AMD Family10h rev B0 and later)
- */
-static void setup_ibs(void)
+/* uninitialize the APIC for the IBS interrupts if needed */
+static void clear_ibs_nmi(void)
 {
-	ibs_allowed = boot_cpu_has(X86_FEATURE_IBS);
+	if (has_ibs)
+		on_each_cpu(apic_clear_ibs_nmi_per_cpu, NULL, 1);
+}
+
+/* initialize the APIC for the IBS interrupts if available */
+static void ibs_init(void)
+{
+	has_ibs = boot_cpu_has(X86_FEATURE_IBS);
 
-	if (!ibs_allowed)
+	if (!has_ibs)
 		return;
 
-	if (pfm_amd64_setup_eilvt()) {
-		ibs_allowed = 0;
+	if (init_ibs_nmi()) {
+		has_ibs = 0;
 		return;
 	}
 
 	printk(KERN_INFO "oprofile: AMD IBS detected\n");
 }
 
-
-/*
- * unitialize the APIC for the IBS interrupts if needed on AMD Family10h
- * rev B0 and later */
-static void clear_ibs_nmi(void)
+static void ibs_exit(void)
 {
-	if (ibs_allowed)
-		on_each_cpu(apic_clear_ibs_nmi_per_cpu, NULL, 1);
+	if (!has_ibs)
+		return;
+
+	clear_ibs_nmi();
 }
 
 static int (*create_arch_files)(struct super_block *sb, struct dentry *root);
@@ -486,7 +433,7 @@ static int setup_ibs_files(struct super_block *sb, struct dentry *root)
 	if (ret)
 		return ret;
 
-	if (!ibs_allowed)
+	if (!has_ibs)
 		return ret;
 
 	/* model specific files */
@@ -519,7 +466,7 @@ static int setup_ibs_files(struct super_block *sb, struct dentry *root)
 
 static int op_amd_init(struct oprofile_operations *ops)
 {
-	setup_ibs();
+	ibs_init();
 	create_arch_files = ops->create_files;
 	ops->create_files = setup_ibs_files;
 	return 0;
@@ -527,10 +474,21 @@ static int op_amd_init(struct oprofile_operations *ops)
 
 static void op_amd_exit(void)
 {
-	clear_ibs_nmi();
+	ibs_exit();
 }
 
-#endif
+#else
+
+/* no IBS support */
+
+static int op_amd_init(struct oprofile_operations *ops)
+{
+	return 0;
+}
+
+static void op_amd_exit(void) {}
+
+#endif /* CONFIG_OPROFILE_IBS */
 
 struct op_x86_model_spec const op_amd_spec = {
 	.init			= op_amd_init,
diff --git a/drivers/oprofile/buffer_sync.c b/drivers/oprofile/buffer_sync.c
index b55cd23..ac014cb 100644
--- a/drivers/oprofile/buffer_sync.c
+++ b/drivers/oprofile/buffer_sync.c
@@ -1,11 +1,12 @@
 /**
  * @file buffer_sync.c
  *
- * @remark Copyright 2002 OProfile authors
+ * @remark Copyright 2002-2009 OProfile authors
  * @remark Read the file COPYING
  *
  * @author John Levon <levon@movementarian.org>
  * @author Barry Kasindorf
+ * @author Robert Richter <robert.richter@amd.com>
  *
  * This is the core of the buffer management. Each
  * CPU buffer is processed and entered into the
@@ -268,18 +269,6 @@ lookup_dcookie(struct mm_struct *mm, unsigned long addr, off_t *offset)
 	return cookie;
 }
 
-static void increment_tail(struct oprofile_cpu_buffer *b)
-{
-	unsigned long new_tail = b->tail_pos + 1;
-
-	rmb();	/* be sure fifo pointers are synchromized */
-
-	if (new_tail < b->buffer_size)
-		b->tail_pos = new_tail;
-	else
-		b->tail_pos = 0;
-}
-
 static unsigned long last_cookie = INVALID_COOKIE;
 
 static void add_cpu_switch(int i)
@@ -327,84 +316,73 @@ static void add_trace_begin(void)
 	add_event_entry(TRACE_BEGIN_CODE);
 }
 
-#ifdef CONFIG_OPROFILE_IBS
-
-#define IBS_FETCH_CODE_SIZE	2
-#define IBS_OP_CODE_SIZE	5
-#define IBS_EIP(offset)				\
-	(((struct op_sample *)&cpu_buf->buffer[(offset)])->eip)
-#define IBS_EVENT(offset)				\
-	(((struct op_sample *)&cpu_buf->buffer[(offset)])->event)
-
-/*
- * Add IBS fetch and op entries to event buffer
- */
-static void add_ibs_begin(struct oprofile_cpu_buffer *cpu_buf, int code,
-			  struct mm_struct *mm)
+static void add_data(struct op_entry *entry, struct mm_struct *mm)
 {
-	unsigned long rip;
-	int i, count;
-	unsigned long ibs_cookie = 0;
+	unsigned long code, pc, val;
+	unsigned long cookie;
 	off_t offset;
 
-	increment_tail(cpu_buf);	/* move to RIP entry */
-
-	rip = IBS_EIP(cpu_buf->tail_pos);
-
-#ifdef __LP64__
-	rip += IBS_EVENT(cpu_buf->tail_pos) << 32;
-#endif
+	if (!op_cpu_buffer_get_data(entry, &code))
+		return;
+	if (!op_cpu_buffer_get_data(entry, &pc))
+		return;
+	if (!op_cpu_buffer_get_size(entry))
+		return;
 
 	if (mm) {
-		ibs_cookie = lookup_dcookie(mm, rip, &offset);
+		cookie = lookup_dcookie(mm, pc, &offset);
 
-		if (ibs_cookie == NO_COOKIE)
-			offset = rip;
-		if (ibs_cookie == INVALID_COOKIE) {
+		if (cookie == NO_COOKIE)
+			offset = pc;
+		if (cookie == INVALID_COOKIE) {
 			atomic_inc(&oprofile_stats.sample_lost_no_mapping);
-			offset = rip;
+			offset = pc;
 		}
-		if (ibs_cookie != last_cookie) {
-			add_cookie_switch(ibs_cookie);
-			last_cookie = ibs_cookie;
+		if (cookie != last_cookie) {
+			add_cookie_switch(cookie);
+			last_cookie = cookie;
 		}
 	} else
-		offset = rip;
+		offset = pc;
 
 	add_event_entry(ESCAPE_CODE);
 	add_event_entry(code);
 	add_event_entry(offset);	/* Offset from Dcookie */
 
-	/* we send the Dcookie offset, but send the raw Linear Add also*/
-	add_event_entry(IBS_EIP(cpu_buf->tail_pos));
-	add_event_entry(IBS_EVENT(cpu_buf->tail_pos));
-
-	if (code == IBS_FETCH_CODE)
-		count = IBS_FETCH_CODE_SIZE;	/*IBS FETCH is 2 int64s*/
-	else
-		count = IBS_OP_CODE_SIZE;	/*IBS OP is 5 int64s*/
-
-	for (i = 0; i < count; i++) {
-		increment_tail(cpu_buf);
-		add_event_entry(IBS_EIP(cpu_buf->tail_pos));
-		add_event_entry(IBS_EVENT(cpu_buf->tail_pos));
-	}
+	while (op_cpu_buffer_get_data(entry, &val))
+		add_event_entry(val);
 }
 
-#endif
-
-static void add_sample_entry(unsigned long offset, unsigned long event)
+static inline void add_sample_entry(unsigned long offset, unsigned long event)
 {
 	add_event_entry(offset);
 	add_event_entry(event);
 }
 
 
-static int add_us_sample(struct mm_struct *mm, struct op_sample *s)
+/*
+ * Add a sample to the global event buffer. If possible the
+ * sample is converted into a persistent dentry/offset pair
+ * for later lookup from userspace. Return 0 on failure.
+ */
+static int
+add_sample(struct mm_struct *mm, struct op_sample *s, int in_kernel)
 {
 	unsigned long cookie;
 	off_t offset;
 
+	if (in_kernel) {
+		add_sample_entry(s->eip, s->event);
+		return 1;
+	}
+
+	/* add userspace sample */
+
+	if (!mm) {
+		atomic_inc(&oprofile_stats.sample_lost_no_mm);
+		return 0;
+	}
+
 	cookie = lookup_dcookie(mm, s->eip, &offset);
 
 	if (cookie == INVALID_COOKIE) {
@@ -423,25 +401,6 @@ static int add_us_sample(struct mm_struct *mm, struct op_sample *s)
 }
 
 
-/* Add a sample to the global event buffer. If possible the
- * sample is converted into a persistent dentry/offset pair
- * for later lookup from userspace.
- */
-static int
-add_sample(struct mm_struct *mm, struct op_sample *s, int in_kernel)
-{
-	if (in_kernel) {
-		add_sample_entry(s->eip, s->event);
-		return 1;
-	} else if (mm) {
-		return add_us_sample(mm, s);
-	} else {
-		atomic_inc(&oprofile_stats.sample_lost_no_mm);
-	}
-	return 0;
-}
-
-
 static void release_mm(struct mm_struct *mm)
 {
 	if (!mm)
@@ -466,33 +425,6 @@ static inline int is_code(unsigned long val)
 }
 
 
-/* "acquire" as many cpu buffer slots as we can */
-static unsigned long get_slots(struct oprofile_cpu_buffer *b)
-{
-	unsigned long head = b->head_pos;
-	unsigned long tail = b->tail_pos;
-
-	/*
-	 * Subtle. This resets the persistent last_task
-	 * and in_kernel values used for switching notes.
-	 * BUT, there is a small window between reading
-	 * head_pos, and this call, that means samples
-	 * can appear at the new head position, but not
-	 * be prefixed with the notes for switching
-	 * kernel mode or a task switch. This small hole
-	 * can lead to mis-attribution or samples where
-	 * we don't know if it's in the kernel or not,
-	 * at the start of an event buffer.
-	 */
-	cpu_buffer_reset(b);
-
-	if (head >= tail)
-		return head - tail;
-
-	return head + (b->buffer_size - tail);
-}
-
-
 /* Move tasks along towards death. Any tasks on dead_tasks
  * will definitely have no remaining references in any
  * CPU buffers at this point, because we use two lists,
@@ -559,71 +491,72 @@ typedef enum {
  */
 void sync_buffer(int cpu)
 {
-	struct oprofile_cpu_buffer *cpu_buf = &per_cpu(cpu_buffer, cpu);
 	struct mm_struct *mm = NULL;
+	struct mm_struct *oldmm;
+	unsigned long val;
 	struct task_struct *new;
 	unsigned long cookie = 0;
 	int in_kernel = 1;
 	sync_buffer_state state = sb_buffer_start;
-#ifndef CONFIG_OPROFILE_IBS
 	unsigned int i;
 	unsigned long available;
-#endif
+	unsigned long flags;
+	struct op_entry entry;
+	struct op_sample *sample;
 
 	mutex_lock(&buffer_mutex);
 
 	add_cpu_switch(cpu);
 
-	/* Remember, only we can modify tail_pos */
-
-#ifndef CONFIG_OPROFILE_IBS
-	available = get_slots(cpu_buf);
+	op_cpu_buffer_reset(cpu);
+	available = op_cpu_buffer_entries(cpu);
 
 	for (i = 0; i < available; ++i) {
-#else
-	while (get_slots(cpu_buf)) {
-#endif
-		struct op_sample *s = &cpu_buf->buffer[cpu_buf->tail_pos];
+		sample = op_cpu_buffer_read_entry(&entry, cpu);
+		if (!sample)
+			break;
 
-		if (is_code(s->eip)) {
-			if (s->event <= CPU_IS_KERNEL) {
+		if (is_code(sample->eip)) {
+			flags = sample->event;
+			if (flags & TRACE_BEGIN) {
+				state = sb_bt_start;
+				add_trace_begin();
+			}
+			if (flags & KERNEL_CTX_SWITCH) {
 				/* kernel/userspace switch */
-				in_kernel = s->event;
+				in_kernel = flags & IS_KERNEL;
 				if (state == sb_buffer_start)
 					state = sb_sample_start;
-				add_kernel_ctx_switch(s->event);
-			} else if (s->event == CPU_TRACE_BEGIN) {
-				state = sb_bt_start;
-				add_trace_begin();
-#ifdef CONFIG_OPROFILE_IBS
-			} else if (s->event == IBS_FETCH_BEGIN) {
-				state = sb_bt_start;
-				add_ibs_begin(cpu_buf, IBS_FETCH_CODE, mm);
-			} else if (s->event == IBS_OP_BEGIN) {
-				state = sb_bt_start;
-				add_ibs_begin(cpu_buf, IBS_OP_CODE, mm);
-#endif
-			} else {
-				struct mm_struct *oldmm = mm;
-
+				add_kernel_ctx_switch(flags & IS_KERNEL);
+			}
+			if (flags & USER_CTX_SWITCH
+			    && op_cpu_buffer_get_data(&entry, &val)) {
 				/* userspace context switch */
-				new = (struct task_struct *)s->event;
-
+				new = (struct task_struct *)val;
+				oldmm = mm;
 				release_mm(oldmm);
 				mm = take_tasks_mm(new);
 				if (mm != oldmm)
 					cookie = get_exec_dcookie(mm);
 				add_user_ctx_switch(new, cookie);
 			}
-		} else if (state >= sb_bt_start &&
-			   !add_sample(mm, s, in_kernel)) {
-			if (state == sb_bt_start) {
-				state = sb_bt_ignore;
-				atomic_inc(&oprofile_stats.bt_lost_no_mapping);
-			}
+			if (op_cpu_buffer_get_size(&entry))
+				add_data(&entry, mm);
+			continue;
 		}
 
-		increment_tail(cpu_buf);
+		if (state < sb_bt_start)
+			/* ignore sample */
+			continue;
+
+		if (add_sample(mm, sample, in_kernel))
+			continue;
+
+		/* ignore backtraces if failed to add a sample */
+		if (state == sb_bt_start) {
+			state = sb_bt_ignore;
+			atomic_inc(&oprofile_stats.bt_lost_no_mapping);
+		}
 	}
 	release_mm(mm);
 
diff --git a/drivers/oprofile/cpu_buffer.c b/drivers/oprofile/cpu_buffer.c
index 01d38e7..2e03b6d 100644
--- a/drivers/oprofile/cpu_buffer.c
+++ b/drivers/oprofile/cpu_buffer.c
@@ -1,11 +1,12 @@
 /**
  * @file cpu_buffer.c
  *
- * @remark Copyright 2002 OProfile authors
+ * @remark Copyright 2002-2009 OProfile authors
  * @remark Read the file COPYING
  *
  * @author John Levon <levon@movementarian.org>
  * @author Barry Kasindorf <barry.kasindorf@amd.com>
+ * @author Robert Richter <robert.richter@amd.com>
  *
  * Each CPU has a local buffer that stores PC value/event
  * pairs. We also log context switches when we notice them.
@@ -28,6 +29,25 @@
 #include "buffer_sync.h"
 #include "oprof.h"
 
+#define OP_BUFFER_FLAGS	0
+
+/*
+ * Read and write access is using spin locking. Thus, writing to the
+ * buffer by NMI handler (x86) could occur also during critical
+ * sections when reading the buffer. To avoid this, there are 2
+ * buffers for independent read and write access. Read access is in
+ * process context only, write access only in the NMI handler. If the
+ * read buffer runs empty, both buffers are swapped atomically. There
+ * is potentially a small window during swapping where the buffers are
+ * disabled and samples could be lost.
+ *
+ * Using 2 buffers is a little bit overhead, but the solution is clear
+ * and does not require changes in the ring buffer implementation. It
+ * can be changed to a single buffer solution when the ring buffer
+ * access is implemented as non-locking atomic code.
+ */
+static struct ring_buffer *op_ring_buffer_read;
+static struct ring_buffer *op_ring_buffer_write;
 DEFINE_PER_CPU(struct oprofile_cpu_buffer, cpu_buffer);
 
 static void wq_sync_buffer(struct work_struct *work);
@@ -35,19 +55,9 @@ static void wq_sync_buffer(struct work_struct *work);
 #define DEFAULT_TIMER_EXPIRE (HZ / 10)
 static int work_enabled;
 
-void free_cpu_buffers(void)
-{
-	int i;
-
-	for_each_possible_cpu(i) {
-		vfree(per_cpu(cpu_buffer, i).buffer);
-		per_cpu(cpu_buffer, i).buffer = NULL;
-	}
-}
-
 unsigned long oprofile_get_cpu_buffer_size(void)
 {
-	return fs_cpu_buffer_size;
+	return oprofile_cpu_buffer_size;
 }
 
 void oprofile_cpu_buffer_inc_smpl_lost(void)
@@ -58,26 +68,36 @@ void oprofile_cpu_buffer_inc_smpl_lost(void)
 	cpu_buf->sample_lost_overflow++;
 }
 
+void free_cpu_buffers(void)
+{
+	if (op_ring_buffer_read)
+		ring_buffer_free(op_ring_buffer_read);
+	op_ring_buffer_read = NULL;
+	if (op_ring_buffer_write)
+		ring_buffer_free(op_ring_buffer_write);
+	op_ring_buffer_write = NULL;
+}
+
 int alloc_cpu_buffers(void)
 {
 	int i;
 
-	unsigned long buffer_size = fs_cpu_buffer_size;
+	unsigned long buffer_size = oprofile_cpu_buffer_size;
+
+	op_ring_buffer_read = ring_buffer_alloc(buffer_size, OP_BUFFER_FLAGS);
+	if (!op_ring_buffer_read)
+		goto fail;
+	op_ring_buffer_write = ring_buffer_alloc(buffer_size, OP_BUFFER_FLAGS);
+	if (!op_ring_buffer_write)
+		goto fail;
 
 	for_each_possible_cpu(i) {
 		struct oprofile_cpu_buffer *b = &per_cpu(cpu_buffer, i);
 
-		b->buffer = vmalloc_node(sizeof(struct op_sample) * buffer_size,
-			cpu_to_node(i));
-		if (!b->buffer)
-			goto fail;
-
 		b->last_task = NULL;
 		b->last_is_kernel = -1;
 		b->tracing = 0;
 		b->buffer_size = buffer_size;
-		b->tail_pos = 0;
-		b->head_pos = 0;
 		b->sample_received = 0;
 		b->sample_lost_overflow = 0;
 		b->backtrace_aborted = 0;
@@ -124,73 +144,156 @@ void end_cpu_work(void)
 	flush_scheduled_work();
 }
 
-/* Resets the cpu buffer to a sane state. */
-void cpu_buffer_reset(struct oprofile_cpu_buffer *cpu_buf)
+/*
+ * This function prepares the cpu buffer to write a sample.
+ *
+ * Struct op_entry is used during operations on the ring buffer while
+ * struct op_sample contains the data that is stored in the ring
+ * buffer. Struct entry can be uninitialized. The function reserves a
+ * data array that is specified by size. Use
+ * op_cpu_buffer_write_commit() after preparing the sample. In case of
+ * errors a null pointer is returned, otherwise the pointer to the
+ * sample.
+ *
+ */
+struct op_sample
+*op_cpu_buffer_write_reserve(struct op_entry *entry, unsigned long size)
 {
-	/* reset these to invalid values; the next sample
-	 * collected will populate the buffer with proper
-	 * values to initialize the buffer
-	 */
-	cpu_buf->last_is_kernel = -1;
-	cpu_buf->last_task = NULL;
+	entry->event = ring_buffer_lock_reserve
+		(op_ring_buffer_write, sizeof(struct op_sample) +
+		 size * sizeof(entry->sample->data[0]), &entry->irq_flags);
+	if (entry->event)
+		entry->sample = ring_buffer_event_data(entry->event);
+	else
+		entry->sample = NULL;
+
+	if (!entry->sample)
+		return NULL;
+
+	entry->size = size;
+	entry->data = entry->sample->data;
+
+	return entry->sample;
 }
 
-/* compute number of available slots in cpu_buffer queue */
-static unsigned long nr_available_slots(struct oprofile_cpu_buffer const *b)
+int op_cpu_buffer_write_commit(struct op_entry *entry)
 {
-	unsigned long head = b->head_pos;
-	unsigned long tail = b->tail_pos;
+	return ring_buffer_unlock_commit(op_ring_buffer_write, entry->event,
+					 entry->irq_flags);
+}
 
-	if (tail > head)
-		return (tail - head) - 1;
+struct op_sample *op_cpu_buffer_read_entry(struct op_entry *entry, int cpu)
+{
+	struct ring_buffer_event *e;
+	e = ring_buffer_consume(op_ring_buffer_read, cpu, NULL);
+	if (e)
+		goto event;
+	if (ring_buffer_swap_cpu(op_ring_buffer_read,
+				 op_ring_buffer_write,
+				 cpu))
+		return NULL;
+	e = ring_buffer_consume(op_ring_buffer_read, cpu, NULL);
+	if (e)
+		goto event;
+	return NULL;
+
+event:
+	entry->event = e;
+	entry->sample = ring_buffer_event_data(e);
+	entry->size = (ring_buffer_event_length(e) - sizeof(struct op_sample))
+		/ sizeof(entry->sample->data[0]);
+	entry->data = entry->sample->data;
+	return entry->sample;
+}
 
-	return tail + (b->buffer_size - head) - 1;
+unsigned long op_cpu_buffer_entries(int cpu)
+{
+	return ring_buffer_entries_cpu(op_ring_buffer_read, cpu)
+		+ ring_buffer_entries_cpu(op_ring_buffer_write, cpu);
 }
 
-static void increment_head(struct oprofile_cpu_buffer *b)
+static int
+op_add_code(struct oprofile_cpu_buffer *cpu_buf, unsigned long backtrace,
+	    int is_kernel, struct task_struct *task)
 {
-	unsigned long new_head = b->head_pos + 1;
+	struct op_entry entry;
+	struct op_sample *sample;
+	unsigned long flags;
+	int size;
+
+	flags = 0;
 
-	/* Ensure anything written to the slot before we
-	 * increment is visible */
-	wmb();
+	if (backtrace)
+		flags |= TRACE_BEGIN;
+
+	/* notice a switch from user->kernel or vice versa */
+	is_kernel = !!is_kernel;
+	if (cpu_buf->last_is_kernel != is_kernel) {
+		cpu_buf->last_is_kernel = is_kernel;
+		flags |= KERNEL_CTX_SWITCH;
+		if (is_kernel)
+			flags |= IS_KERNEL;
+	}
+
+	/* notice a task switch */
+	if (cpu_buf->last_task != task) {
+		cpu_buf->last_task = task;
+		flags |= USER_CTX_SWITCH;
+	}
+
+	if (!flags)
+		/* nothing to do */
+		return 0;
 
-	if (new_head < b->buffer_size)
-		b->head_pos = new_head;
+	if (flags & USER_CTX_SWITCH)
+		size = 1;
 	else
-		b->head_pos = 0;
-}
+		size = 0;
 
-static inline void
-add_sample(struct oprofile_cpu_buffer *cpu_buf,
-	   unsigned long pc, unsigned long event)
-{
-	struct op_sample *entry = &cpu_buf->buffer[cpu_buf->head_pos];
-	entry->eip = pc;
-	entry->event = event;
-	increment_head(cpu_buf);
+	sample = op_cpu_buffer_write_reserve(&entry, size);
+	if (!sample)
+		return -ENOMEM;
+
+	sample->eip = ESCAPE_CODE;
+	sample->event = flags;
+
+	if (size)
+		op_cpu_buffer_add_data(&entry, (unsigned long)task);
+
+	op_cpu_buffer_write_commit(&entry);
+
+	return 0;
 }
 
-static inline void
-add_code(struct oprofile_cpu_buffer *buffer, unsigned long value)
+static inline int
+op_add_sample(struct oprofile_cpu_buffer *cpu_buf,
+	      unsigned long pc, unsigned long event)
 {
-	add_sample(buffer, ESCAPE_CODE, value);
+	struct op_entry entry;
+	struct op_sample *sample;
+
+	sample = op_cpu_buffer_write_reserve(&entry, 0);
+	if (!sample)
+		return -ENOMEM;
+
+	sample->eip = pc;
+	sample->event = event;
+
+	return op_cpu_buffer_write_commit(&entry);
 }
 
-/* This must be safe from any context. It's safe writing here
- * because of the head/tail separation of the writer and reader
- * of the CPU buffer.
+/*
+ * This must be safe from any context.
  *
  * is_kernel is needed because on some architectures you cannot
  * tell if you are in kernel or user space simply by looking at
  * pc. We tag this in the buffer by generating kernel enter/exit
  * events whenever is_kernel changes
  */
-static int log_sample(struct oprofile_cpu_buffer *cpu_buf, unsigned long pc,
-		      int is_kernel, unsigned long event)
+static int
+log_sample(struct oprofile_cpu_buffer *cpu_buf, unsigned long pc,
+	   unsigned long backtrace, int is_kernel, unsigned long event)
 {
-	struct task_struct *task;
-
 	cpu_buf->sample_received++;
 
 	if (pc == ESCAPE_CODE) {
@@ -198,131 +301,115 @@ static int log_sample(struct oprofile_cpu_buffer *cpu_buf, unsigned long pc,
 		return 0;
 	}
 
-	if (nr_available_slots(cpu_buf) < 3) {
-		cpu_buf->sample_lost_overflow++;
-		return 0;
-	}
-
-	is_kernel = !!is_kernel;
+	if (op_add_code(cpu_buf, backtrace, is_kernel, current))
+		goto fail;
 
-	task = current;
+	if (op_add_sample(cpu_buf, pc, event))
+		goto fail;
 
-	/* notice a switch from user->kernel or vice versa */
-	if (cpu_buf->last_is_kernel != is_kernel) {
-		cpu_buf->last_is_kernel = is_kernel;
-		add_code(cpu_buf, is_kernel);
-	}
-
-	/* notice a task switch */
-	if (cpu_buf->last_task != task) {
-		cpu_buf->last_task = task;
-		add_code(cpu_buf, (unsigned long)task);
-	}
-
-	add_sample(cpu_buf, pc, event);
 	return 1;
+
+fail:
+	cpu_buf->sample_lost_overflow++;
+	return 0;
 }
 
-static int oprofile_begin_trace(struct oprofile_cpu_buffer *cpu_buf)
+static inline void oprofile_begin_trace(struct oprofile_cpu_buffer *cpu_buf)
 {
-	if (nr_available_slots(cpu_buf) < 4) {
-		cpu_buf->sample_lost_overflow++;
-		return 0;
-	}
-
-	add_code(cpu_buf, CPU_TRACE_BEGIN);
 	cpu_buf->tracing = 1;
-	return 1;
 }
 
-static void oprofile_end_trace(struct oprofile_cpu_buffer *cpu_buf)
+static inline void oprofile_end_trace(struct oprofile_cpu_buffer *cpu_buf)
 {
 	cpu_buf->tracing = 0;
 }
 
-void oprofile_add_ext_sample(unsigned long pc, struct pt_regs * const regs,
-				unsigned long event, int is_kernel)
+static inline void
+__oprofile_add_ext_sample(unsigned long pc, struct pt_regs * const regs,
+			  unsigned long event, int is_kernel)
 {
 	struct oprofile_cpu_buffer *cpu_buf = &__get_cpu_var(cpu_buffer);
+	unsigned long backtrace = oprofile_backtrace_depth;
 
-	if (!backtrace_depth) {
-		log_sample(cpu_buf, pc, is_kernel, event);
+	/*
+	 * if log_sample() fail we can't backtrace since we lost the
+	 * source of this event
+	 */
+	if (!log_sample(cpu_buf, pc, backtrace, is_kernel, event))
+		/* failed */
 		return;
-	}
 
-	if (!oprofile_begin_trace(cpu_buf))
+	if (!backtrace)
 		return;
 
-	/* if log_sample() fail we can't backtrace since we lost the source
-	 * of this event */
-	if (log_sample(cpu_buf, pc, is_kernel, event))
-		oprofile_ops.backtrace(regs, backtrace_depth);
+	oprofile_begin_trace(cpu_buf);
+	oprofile_ops.backtrace(regs, backtrace);
 	oprofile_end_trace(cpu_buf);
 }
 
+void oprofile_add_ext_sample(unsigned long pc, struct pt_regs * const regs,
+			     unsigned long event, int is_kernel)
+{
+	__oprofile_add_ext_sample(pc, regs, event, is_kernel);
+}
+
 void oprofile_add_sample(struct pt_regs * const regs, unsigned long event)
 {
 	int is_kernel = !user_mode(regs);
 	unsigned long pc = profile_pc(regs);
 
-	oprofile_add_ext_sample(pc, regs, event, is_kernel);
+	__oprofile_add_ext_sample(pc, regs, event, is_kernel);
 }
 
-#ifdef CONFIG_OPROFILE_IBS
-
-#define MAX_IBS_SAMPLE_SIZE 14
-
-void oprofile_add_ibs_sample(struct pt_regs *const regs,
-			     unsigned int *const ibs_sample, int ibs_code)
+/*
+ * Add samples with data to the ring buffer.
+ *
+ * Use oprofile_add_data(&entry, val) to add data and
+ * oprofile_write_commit(&entry) to commit the sample.
+ */
+void
+oprofile_write_reserve(struct op_entry *entry, struct pt_regs * const regs,
+		       unsigned long pc, int code, int size)
 {
+	struct op_sample *sample;
 	int is_kernel = !user_mode(regs);
 	struct oprofile_cpu_buffer *cpu_buf = &__get_cpu_var(cpu_buffer);
-	struct task_struct *task;
 
 	cpu_buf->sample_received++;
 
-	if (nr_available_slots(cpu_buf) < MAX_IBS_SAMPLE_SIZE) {
-		/* we can't backtrace since we lost the source of this event */
-		cpu_buf->sample_lost_overflow++;
-		return;
-	}
+	/* no backtraces for samples with data */
+	if (op_add_code(cpu_buf, 0, is_kernel, current))
+		goto fail;
 
-	/* notice a switch from user->kernel or vice versa */
-	if (cpu_buf->last_is_kernel != is_kernel) {
-		cpu_buf->last_is_kernel = is_kernel;
-		add_code(cpu_buf, is_kernel);
-	}
+	sample = op_cpu_buffer_write_reserve(entry, size + 2);
+	if (!sample)
+		goto fail;
+	sample->eip = ESCAPE_CODE;
+	sample->event = 0;		/* no flags */
 
-	/* notice a task switch */
-	if (!is_kernel) {
-		task = current;
-		if (cpu_buf->last_task != task) {
-			cpu_buf->last_task = task;
-			add_code(cpu_buf, (unsigned long)task);
-		}
-	}
+	op_cpu_buffer_add_data(entry, code);
+	op_cpu_buffer_add_data(entry, pc);
 
-	add_code(cpu_buf, ibs_code);
-	add_sample(cpu_buf, ibs_sample[0], ibs_sample[1]);
-	add_sample(cpu_buf, ibs_sample[2], ibs_sample[3]);
-	add_sample(cpu_buf, ibs_sample[4], ibs_sample[5]);
+	return;
 
-	if (ibs_code == IBS_OP_BEGIN) {
-		add_sample(cpu_buf, ibs_sample[6], ibs_sample[7]);
-		add_sample(cpu_buf, ibs_sample[8], ibs_sample[9]);
-		add_sample(cpu_buf, ibs_sample[10], ibs_sample[11]);
-	}
+fail:
+	cpu_buf->sample_lost_overflow++;
+}
 
-	if (backtrace_depth)
-		oprofile_ops.backtrace(regs, backtrace_depth);
+int oprofile_add_data(struct op_entry *entry, unsigned long val)
+{
+	return op_cpu_buffer_add_data(entry, val);
 }
 
-#endif
+int oprofile_write_commit(struct op_entry *entry)
+{
+	return op_cpu_buffer_write_commit(entry);
+}
 
 void oprofile_add_pc(unsigned long pc, int is_kernel, unsigned long event)
 {
 	struct oprofile_cpu_buffer *cpu_buf = &__get_cpu_var(cpu_buffer);
-	log_sample(cpu_buf, pc, is_kernel, event);
+	log_sample(cpu_buf, pc, 0, is_kernel, event);
 }
 
 void oprofile_add_trace(unsigned long pc)
@@ -332,21 +419,21 @@ void oprofile_add_trace(unsigned long pc)
 	if (!cpu_buf->tracing)
 		return;
 
-	if (nr_available_slots(cpu_buf) < 1) {
-		cpu_buf->tracing = 0;
-		cpu_buf->sample_lost_overflow++;
-		return;
-	}
+	/*
+	 * broken frame can give an eip with the same value as an
+	 * escape code, abort the trace if we get it
+	 */
+	if (pc == ESCAPE_CODE)
+		goto fail;
 
-	/* broken frame can give an eip with the same value as an escape code,
-	 * abort the trace if we get it */
-	if (pc == ESCAPE_CODE) {
-		cpu_buf->tracing = 0;
-		cpu_buf->backtrace_aborted++;
-		return;
-	}
+	if (op_add_sample(cpu_buf, pc, 0))
+		goto fail;
 
-	add_sample(cpu_buf, pc, 0);
+	return;
+fail:
+	cpu_buf->tracing = 0;
+	cpu_buf->backtrace_aborted++;
+	return;
 }
 
 /*
diff --git a/drivers/oprofile/cpu_buffer.h b/drivers/oprofile/cpu_buffer.h
index d3cc262..63f81c4 100644
--- a/drivers/oprofile/cpu_buffer.h
+++ b/drivers/oprofile/cpu_buffer.h
@@ -1,10 +1,11 @@
 /**
  * @file cpu_buffer.h
  *
- * @remark Copyright 2002 OProfile authors
+ * @remark Copyright 2002-2009 OProfile authors
  * @remark Read the file COPYING
  *
  * @author John Levon <levon@movementarian.org>
+ * @author Robert Richter <robert.richter@amd.com>
  */
 
 #ifndef OPROFILE_CPU_BUFFER_H
@@ -15,6 +16,7 @@
 #include <linux/workqueue.h>
 #include <linux/cache.h>
 #include <linux/sched.h>
+#include <linux/ring_buffer.h>
 
 struct task_struct;
 
@@ -30,16 +32,16 @@ void end_cpu_work(void);
 struct op_sample {
 	unsigned long eip;
 	unsigned long event;
+	unsigned long data[0];
 };
 
+struct op_entry;
+
 struct oprofile_cpu_buffer {
-	volatile unsigned long head_pos;
-	volatile unsigned long tail_pos;
 	unsigned long buffer_size;
 	struct task_struct *last_task;
 	int last_is_kernel;
 	int tracing;
-	struct op_sample *buffer;
 	unsigned long sample_received;
 	unsigned long sample_lost_overflow;
 	unsigned long backtrace_aborted;
@@ -50,12 +52,62 @@ struct oprofile_cpu_buffer {
 
 DECLARE_PER_CPU(struct oprofile_cpu_buffer, cpu_buffer);
 
-void cpu_buffer_reset(struct oprofile_cpu_buffer *cpu_buf);
+/*
+ * Resets the cpu buffer to a sane state.
+ *
+ * reset these to invalid values; the next sample collected will
+ * populate the buffer with proper values to initialize the buffer
+ */
+static inline void op_cpu_buffer_reset(int cpu)
+{
+	struct oprofile_cpu_buffer *cpu_buf = &per_cpu(cpu_buffer, cpu);
+
+	cpu_buf->last_is_kernel = -1;
+	cpu_buf->last_task = NULL;
+}
+
+struct op_sample
+*op_cpu_buffer_write_reserve(struct op_entry *entry, unsigned long size);
+int op_cpu_buffer_write_commit(struct op_entry *entry);
+struct op_sample *op_cpu_buffer_read_entry(struct op_entry *entry, int cpu);
+unsigned long op_cpu_buffer_entries(int cpu);
+
+/* returns the remaining free size of data in the entry */
+static inline
+int op_cpu_buffer_add_data(struct op_entry *entry, unsigned long val)
+{
+	if (!entry->size)
+		return 0;
+	*entry->data = val;
+	entry->size--;
+	entry->data++;
+	return entry->size;
+}
+
+/* returns the size of data in the entry */
+static inline
+int op_cpu_buffer_get_size(struct op_entry *entry)
+{
+	return entry->size;
+}
+
+/* returns 0 if empty or the size of data including the current value */
+static inline
+int op_cpu_buffer_get_data(struct op_entry *entry, unsigned long *val)
+{
+	int size = entry->size;
+	if (!size)
+		return 0;
+	*val = *entry->data;
+	entry->size--;
+	entry->data++;
+	return size;
+}
 
-/* transient events for the CPU buffer -> event buffer */
-#define CPU_IS_KERNEL 1
-#define CPU_TRACE_BEGIN 2
-#define IBS_FETCH_BEGIN 3
-#define IBS_OP_BEGIN    4
+/* extra data flags */
+#define KERNEL_CTX_SWITCH	(1UL << 0)
+#define IS_KERNEL		(1UL << 1)
+#define TRACE_BEGIN		(1UL << 2)
+#define USER_CTX_SWITCH		(1UL << 3)
 
 #endif /* OPROFILE_CPU_BUFFER_H */
diff --git a/drivers/oprofile/event_buffer.c b/drivers/oprofile/event_buffer.c
index 191a320..2b7ae36 100644
--- a/drivers/oprofile/event_buffer.c
+++ b/drivers/oprofile/event_buffer.c
@@ -73,8 +73,8 @@ int alloc_event_buffer(void)
 	unsigned long flags;
 
 	spin_lock_irqsave(&oprofilefs_lock, flags);
-	buffer_size = fs_buffer_size;
-	buffer_watershed = fs_buffer_watershed;
+	buffer_size = oprofile_buffer_size;
+	buffer_watershed = oprofile_buffer_watershed;
 	spin_unlock_irqrestore(&oprofilefs_lock, flags);
 
 	if (buffer_watershed >= buffer_size)
diff --git a/drivers/oprofile/oprof.c b/drivers/oprofile/oprof.c
index cd37590..3cffce9 100644
--- a/drivers/oprofile/oprof.c
+++ b/drivers/oprofile/oprof.c
@@ -23,7 +23,7 @@
 struct oprofile_operations oprofile_ops;
 
 unsigned long oprofile_started;
-unsigned long backtrace_depth;
+unsigned long oprofile_backtrace_depth;
 static unsigned long is_setup;
 static DEFINE_MUTEX(start_mutex);
 
@@ -172,7 +172,7 @@ int oprofile_set_backtrace(unsigned long val)
 		goto out;
 	}
 
-	backtrace_depth = val;
+	oprofile_backtrace_depth = val;
 
 out:
 	mutex_unlock(&start_mutex);
diff --git a/drivers/oprofile/oprof.h b/drivers/oprofile/oprof.h
index 5df0c21..c288d3c 100644
--- a/drivers/oprofile/oprof.h
+++ b/drivers/oprofile/oprof.h
@@ -21,12 +21,12 @@ void oprofile_stop(void);
 
 struct oprofile_operations;
 
-extern unsigned long fs_buffer_size;
-extern unsigned long fs_cpu_buffer_size;
-extern unsigned long fs_buffer_watershed;
+extern unsigned long oprofile_buffer_size;
+extern unsigned long oprofile_cpu_buffer_size;
+extern unsigned long oprofile_buffer_watershed;
 extern struct oprofile_operations oprofile_ops;
 extern unsigned long oprofile_started;
-extern unsigned long backtrace_depth;
+extern unsigned long oprofile_backtrace_depth;
 
 struct super_block;
 struct dentry;
diff --git a/drivers/oprofile/oprofile_files.c b/drivers/oprofile/oprofile_files.c
index cc106d5..5d36ffc 100644
--- a/drivers/oprofile/oprofile_files.c
+++ b/drivers/oprofile/oprofile_files.c
@@ -14,13 +14,18 @@
 #include "oprofile_stats.h"
 #include "oprof.h"
 
-unsigned long fs_buffer_size = 131072;
-unsigned long fs_cpu_buffer_size = 8192;
-unsigned long fs_buffer_watershed = 32768; /* FIXME: tune */
+#define BUFFER_SIZE_DEFAULT		131072
+#define CPU_BUFFER_SIZE_DEFAULT		8192
+#define BUFFER_WATERSHED_DEFAULT	32768	/* FIXME: tune */
+
+unsigned long oprofile_buffer_size;
+unsigned long oprofile_cpu_buffer_size;
+unsigned long oprofile_buffer_watershed;
 
 static ssize_t depth_read(struct file *file, char __user *buf, size_t count, loff_t *offset)
 {
-	return oprofilefs_ulong_to_user(backtrace_depth, buf, count, offset);
+	return oprofilefs_ulong_to_user(oprofile_backtrace_depth, buf, count,
+					offset);
 }
 
 
@@ -120,12 +125,17 @@ static const struct file_operations dump_fops = {
 
 void oprofile_create_files(struct super_block *sb, struct dentry *root)
 {
+	/* reinitialize default values */
+	oprofile_buffer_size =		BUFFER_SIZE_DEFAULT;
+	oprofile_cpu_buffer_size =	CPU_BUFFER_SIZE_DEFAULT;
+	oprofile_buffer_watershed =	BUFFER_WATERSHED_DEFAULT;
+
 	oprofilefs_create_file(sb, root, "enable", &enable_fops);
 	oprofilefs_create_file_perm(sb, root, "dump", &dump_fops, 0666);
 	oprofilefs_create_file(sb, root, "buffer", &event_buffer_fops);
-	oprofilefs_create_ulong(sb, root, "buffer_size", &fs_buffer_size);
-	oprofilefs_create_ulong(sb, root, "buffer_watershed", &fs_buffer_watershed);
-	oprofilefs_create_ulong(sb, root, "cpu_buffer_size", &fs_cpu_buffer_size);
+	oprofilefs_create_ulong(sb, root, "buffer_size", &oprofile_buffer_size);
+	oprofilefs_create_ulong(sb, root, "buffer_watershed", &oprofile_buffer_watershed);
+	oprofilefs_create_ulong(sb, root, "cpu_buffer_size", &oprofile_cpu_buffer_size);
 	oprofilefs_create_file(sb, root, "cpu_type", &cpu_type_fops);
 	oprofilefs_create_file(sb, root, "backtrace_depth", &depth_fops);
 	oprofilefs_create_file(sb, root, "pointer_size", &pointer_size_fops);
diff --git a/include/linux/oprofile.h b/include/linux/oprofile.h
index 5231861..1d9518b 100644
--- a/include/linux/oprofile.h
+++ b/include/linux/oprofile.h
@@ -86,8 +86,7 @@ int oprofile_arch_init(struct oprofile_operations * ops);
 void oprofile_arch_exit(void);
 
 /**
- * Add a sample. This may be called from any context. Pass
- * smp_processor_id() as cpu.
+ * Add a sample. This may be called from any context.
  */
 void oprofile_add_sample(struct pt_regs * const regs, unsigned long event);
 
@@ -165,4 +164,22 @@ void oprofile_put_buff(unsigned long *buf, unsigned int start,
 unsigned long oprofile_get_cpu_buffer_size(void);
 void oprofile_cpu_buffer_inc_smpl_lost(void);
  
+/* cpu buffer functions */
+
+struct op_sample;
+
+struct op_entry {
+	struct ring_buffer_event *event;
+	struct op_sample *sample;
+	unsigned long irq_flags;
+	unsigned long size;
+	unsigned long *data;
+};
+
+void oprofile_write_reserve(struct op_entry *entry,
+			    struct pt_regs * const regs,
+			    unsigned long pc, int code, int size);
+int oprofile_add_data(struct op_entry *entry, unsigned long val);
+int oprofile_write_commit(struct op_entry *entry);
+
 #endif /* OPROFILE_H */
diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h
index e097c2e..de9d8c1 100644
--- a/include/linux/ring_buffer.h
+++ b/include/linux/ring_buffer.h
@@ -116,6 +116,8 @@ void ring_buffer_record_enable_cpu(struct ring_buffer *buffer, int cpu);
 
 unsigned long ring_buffer_entries(struct ring_buffer *buffer);
 unsigned long ring_buffer_overruns(struct ring_buffer *buffer);
+unsigned long ring_buffer_entries_cpu(struct ring_buffer *buffer, int cpu);
+unsigned long ring_buffer_overrun_cpu(struct ring_buffer *buffer, int cpu);
 
 u64 ring_buffer_time_stamp(int cpu);
 void ring_buffer_normalize_time_stamp(int cpu, u64 *ts);
diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index 668bbb5..d42b882 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -31,6 +31,7 @@ void tracing_on(void)
 {
 	ring_buffers_off = 0;
 }
+EXPORT_SYMBOL_GPL(tracing_on);
 
 /**
  * tracing_off - turn off all tracing buffers
@@ -44,6 +45,7 @@ void tracing_off(void)
 {
 	ring_buffers_off = 1;
 }
+EXPORT_SYMBOL_GPL(tracing_off);
 
 /* Up this if you want to test the TIME_EXTENTS and normalization */
 #define DEBUG_SHIFT 0
@@ -60,12 +62,14 @@ u64 ring_buffer_time_stamp(int cpu)
 
 	return time;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_time_stamp);
 
 void ring_buffer_normalize_time_stamp(int cpu, u64 *ts)
 {
 	/* Just stupid testing the normalize function and deltas */
 	*ts >>= DEBUG_SHIFT;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_normalize_time_stamp);
 
 #define RB_EVNT_HDR_SIZE (sizeof(struct ring_buffer_event))
 #define RB_ALIGNMENT_SHIFT	2
@@ -113,8 +117,15 @@ rb_event_length(struct ring_buffer_event *event)
  */
 unsigned ring_buffer_event_length(struct ring_buffer_event *event)
 {
-	return rb_event_length(event);
+	unsigned length = rb_event_length(event);
+	if (event->type != RINGBUF_TYPE_DATA)
+		return length;
+	length -= RB_EVNT_HDR_SIZE;
+	if (length > RB_MAX_SMALL_DATA + sizeof(event->array[0]))
+                length -= sizeof(event->array[0]);
+	return length;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_event_length);
 
 /* inline for ring buffer fast paths */
 static inline void *
@@ -136,6 +147,7 @@ void *ring_buffer_event_data(struct ring_buffer_event *event)
 {
 	return rb_event_data(event);
 }
+EXPORT_SYMBOL_GPL(ring_buffer_event_data);
 
 #define for_each_buffer_cpu(buffer, cpu)		\
 	for_each_cpu_mask(cpu, buffer->cpumask)
@@ -381,7 +393,7 @@ extern int ring_buffer_page_too_big(void);
 
 /**
  * ring_buffer_alloc - allocate a new ring_buffer
- * @size: the size in bytes that is needed.
+ * @size: the size in bytes per cpu that is needed.
  * @flags: attributes to set for the ring buffer.
  *
  * Currently the only flag that is available is the RB_FL_OVERWRITE
@@ -444,6 +456,7 @@ struct ring_buffer *ring_buffer_alloc(unsigned long size, unsigned flags)
 	kfree(buffer);
 	return NULL;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_alloc);
 
 /**
  * ring_buffer_free - free a ring buffer.
@@ -459,6 +472,7 @@ ring_buffer_free(struct ring_buffer *buffer)
 
 	kfree(buffer);
 }
+EXPORT_SYMBOL_GPL(ring_buffer_free);
 
 static void rb_reset_cpu(struct ring_buffer_per_cpu *cpu_buffer);
 
@@ -620,6 +634,7 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size)
 	mutex_unlock(&buffer->mutex);
 	return -ENOMEM;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_resize);
 
 static inline int rb_null_event(struct ring_buffer_event *event)
 {
@@ -1220,6 +1235,7 @@ ring_buffer_lock_reserve(struct ring_buffer *buffer,
 		preempt_enable_notrace();
 	return NULL;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_lock_reserve);
 
 static void rb_commit(struct ring_buffer_per_cpu *cpu_buffer,
 		      struct ring_buffer_event *event)
@@ -1269,6 +1285,7 @@ int ring_buffer_unlock_commit(struct ring_buffer *buffer,
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_unlock_commit);
 
 /**
  * ring_buffer_write - write data to the buffer without reserving
@@ -1334,6 +1351,7 @@ int ring_buffer_write(struct ring_buffer *buffer,
 
 	return ret;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_write);
 
 static inline int rb_per_cpu_empty(struct ring_buffer_per_cpu *cpu_buffer)
 {
@@ -1360,6 +1378,7 @@ void ring_buffer_record_disable(struct ring_buffer *buffer)
 {
 	atomic_inc(&buffer->record_disabled);
 }
+EXPORT_SYMBOL_GPL(ring_buffer_record_disable);
 
 /**
  * ring_buffer_record_enable - enable writes to the buffer
@@ -1372,6 +1391,7 @@ void ring_buffer_record_enable(struct ring_buffer *buffer)
 {
 	atomic_dec(&buffer->record_disabled);
 }
+EXPORT_SYMBOL_GPL(ring_buffer_record_enable);
 
 /**
  * ring_buffer_record_disable_cpu - stop all writes into the cpu_buffer
@@ -1393,6 +1413,7 @@ void ring_buffer_record_disable_cpu(struct ring_buffer *buffer, int cpu)
 	cpu_buffer = buffer->buffers[cpu];
 	atomic_inc(&cpu_buffer->record_disabled);
 }
+EXPORT_SYMBOL_GPL(ring_buffer_record_disable_cpu);
 
 /**
  * ring_buffer_record_enable_cpu - enable writes to the buffer
@@ -1412,6 +1433,7 @@ void ring_buffer_record_enable_cpu(struct ring_buffer *buffer, int cpu)
 	cpu_buffer = buffer->buffers[cpu];
 	atomic_dec(&cpu_buffer->record_disabled);
 }
+EXPORT_SYMBOL_GPL(ring_buffer_record_enable_cpu);
 
 /**
  * ring_buffer_entries_cpu - get the number of entries in a cpu buffer
@@ -1428,6 +1450,7 @@ unsigned long ring_buffer_entries_cpu(struct ring_buffer *buffer, int cpu)
 	cpu_buffer = buffer->buffers[cpu];
 	return cpu_buffer->entries;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_entries_cpu);
 
 /**
  * ring_buffer_overrun_cpu - get the number of overruns in a cpu_buffer
@@ -1444,6 +1467,7 @@ unsigned long ring_buffer_overrun_cpu(struct ring_buffer *buffer, int cpu)
 	cpu_buffer = buffer->buffers[cpu];
 	return cpu_buffer->overrun;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_overrun_cpu);
 
 /**
  * ring_buffer_entries - get the number of entries in a buffer
@@ -1466,6 +1490,7 @@ unsigned long ring_buffer_entries(struct ring_buffer *buffer)
 
 	return entries;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_entries);
 
 /**
  * ring_buffer_overrun_cpu - get the number of overruns in buffer
@@ -1488,6 +1513,7 @@ unsigned long ring_buffer_overruns(struct ring_buffer *buffer)
 
 	return overruns;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_overruns);
 
 /**
  * ring_buffer_iter_reset - reset an iterator
@@ -1513,6 +1539,7 @@ void ring_buffer_iter_reset(struct ring_buffer_iter *iter)
 	else
 		iter->read_stamp = iter->head_page->time_stamp;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_iter_reset);
 
 /**
  * ring_buffer_iter_empty - check if an iterator has no more to read
@@ -1527,6 +1554,7 @@ int ring_buffer_iter_empty(struct ring_buffer_iter *iter)
 	return iter->head_page == cpu_buffer->commit_page &&
 		iter->head == rb_commit_index(cpu_buffer);
 }
+EXPORT_SYMBOL_GPL(ring_buffer_iter_empty);
 
 static void
 rb_update_read_stamp(struct ring_buffer_per_cpu *cpu_buffer,
@@ -1797,6 +1825,7 @@ ring_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts)
 
 	return NULL;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_peek);
 
 /**
  * ring_buffer_iter_peek - peek at the next event to be read
@@ -1867,6 +1896,7 @@ ring_buffer_iter_peek(struct ring_buffer_iter *iter, u64 *ts)
 
 	return NULL;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_iter_peek);
 
 /**
  * ring_buffer_consume - return an event and consume it
@@ -1894,6 +1924,7 @@ ring_buffer_consume(struct ring_buffer *buffer, int cpu, u64 *ts)
 
 	return event;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_consume);
 
 /**
  * ring_buffer_read_start - start a non consuming read of the buffer
@@ -1934,6 +1965,7 @@ ring_buffer_read_start(struct ring_buffer *buffer, int cpu)
 
 	return iter;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_read_start);
 
 /**
  * ring_buffer_finish - finish reading the iterator of the buffer
@@ -1950,6 +1982,7 @@ ring_buffer_read_finish(struct ring_buffer_iter *iter)
 	atomic_dec(&cpu_buffer->record_disabled);
 	kfree(iter);
 }
+EXPORT_SYMBOL_GPL(ring_buffer_read_finish);
 
 /**
  * ring_buffer_read - read the next item in the ring buffer by the iterator
@@ -1971,6 +2004,7 @@ ring_buffer_read(struct ring_buffer_iter *iter, u64 *ts)
 
 	return event;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_read);
 
 /**
  * ring_buffer_size - return the size of the ring buffer (in bytes)
@@ -1980,6 +2014,7 @@ unsigned long ring_buffer_size(struct ring_buffer *buffer)
 {
 	return BUF_PAGE_SIZE * buffer->pages;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_size);
 
 static void
 rb_reset_cpu(struct ring_buffer_per_cpu *cpu_buffer)
@@ -2022,6 +2057,7 @@ void ring_buffer_reset_cpu(struct ring_buffer *buffer, int cpu)
 
 	spin_unlock_irqrestore(&cpu_buffer->lock, flags);
 }
+EXPORT_SYMBOL_GPL(ring_buffer_reset_cpu);
 
 /**
  * ring_buffer_reset - reset a ring buffer
@@ -2034,6 +2070,7 @@ void ring_buffer_reset(struct ring_buffer *buffer)
 	for_each_buffer_cpu(buffer, cpu)
 		ring_buffer_reset_cpu(buffer, cpu);
 }
+EXPORT_SYMBOL_GPL(ring_buffer_reset);
 
 /**
  * rind_buffer_empty - is the ring buffer empty?
@@ -2052,6 +2089,7 @@ int ring_buffer_empty(struct ring_buffer *buffer)
 	}
 	return 1;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_empty);
 
 /**
  * ring_buffer_empty_cpu - is a cpu buffer of a ring buffer empty?
@@ -2068,6 +2106,7 @@ int ring_buffer_empty_cpu(struct ring_buffer *buffer, int cpu)
 	cpu_buffer = buffer->buffers[cpu];
 	return rb_per_cpu_empty(cpu_buffer);
 }
+EXPORT_SYMBOL_GPL(ring_buffer_empty_cpu);
 
 /**
  * ring_buffer_swap_cpu - swap a CPU buffer between two ring buffers
@@ -2117,6 +2156,7 @@ int ring_buffer_swap_cpu(struct ring_buffer *buffer_a,
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(ring_buffer_swap_cpu);
 
 static ssize_t
 rb_simple_read(struct file *filp, char __user *ubuf,
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index d86e325..a96b335 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -914,7 +914,7 @@ enum trace_file_type {
 	TRACE_FILE_LAT_FMT	= 1,
 };
 
-static void trace_iterator_increment(struct trace_iterator *iter, int cpu)
+static void trace_iterator_increment(struct trace_iterator *iter)
 {
 	/* Don't allow ftrace to trace into the ring buffers */
 	ftrace_disable_cpu();
@@ -993,7 +993,7 @@ static void *find_next_entry_inc(struct trace_iterator *iter)
 	iter->ent = __find_next_entry(iter, &iter->cpu, &iter->ts);
 
 	if (iter->ent)
-		trace_iterator_increment(iter, iter->cpu);
+		trace_iterator_increment(iter);
 
 	return iter->ent ? iter : NULL;
 }