15 files changed, 992 insertions, 329 deletions

diff --git a/tools/perf/Makefile b/tools/perf/Makefile
index 0cbd5d6874e..36d7eef4991 100644
--- a/tools/perf/Makefile
+++ b/tools/perf/Makefile
@@ -157,10 +157,15 @@ uname_R := $(shell sh -c 'uname -r 2>/dev/null || echo not')
 uname_P := $(shell sh -c 'uname -p 2>/dev/null || echo not')
 uname_V := $(shell sh -c 'uname -v 2>/dev/null || echo not')
 
+# If we're on a 64-bit kernel, use -m64
+ifneq ($(patsubst %64,%,$(uname_M)),$(uname_M))
+  M64 := -m64
+endif
+
 # CFLAGS and LDFLAGS are for the users to override from the command line.
 
-CFLAGS = -ggdb3 -Wall -Werror -Wstrict-prototypes -Wmissing-declarations -Wmissing-prototypes -std=gnu99 -Wdeclaration-after-statement -O6
-LDFLAGS = -lpthread -lrt -lelf
+CFLAGS = $(M64) -ggdb3 -Wall -Wstrict-prototypes -Wmissing-declarations -Wmissing-prototypes -std=gnu99 -Wdeclaration-after-statement -Werror -O6
+LDFLAGS = -lpthread -lrt -lelf -lm
 ALL_CFLAGS = $(CFLAGS)
 ALL_LDFLAGS = $(LDFLAGS)
 STRIP ?= strip
@@ -285,6 +290,7 @@ LIB_FILE=libperf.a
 
 LIB_H += ../../include/linux/perf_counter.h
 LIB_H += perf.h
+LIB_H += types.h
 LIB_H += util/list.h
 LIB_H += util/rbtree.h
 LIB_H += util/levenshtein.h
diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index b1ed5f766cb..7e58e3ad150 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -25,6 +25,10 @@
 #define SHOW_USER	2
 #define SHOW_HV		4
 
+#define MIN_GREEN		0.5
+#define MIN_RED		5.0
+
+
 static char		const *input_name = "perf.data";
 static char		*vmlinux = "vmlinux";
 
@@ -39,40 +43,42 @@ static int		dump_trace = 0;
 
 static int		verbose;
 
+static int		print_line;
+
 static unsigned long	page_size;
 static unsigned long	mmap_window = 32;
 
 struct ip_event {
 	struct perf_event_header header;
-	__u64 ip;
-	__u32 pid, tid;
+	u64 ip;
+	u32 pid, tid;
 };
 
 struct mmap_event {
 	struct perf_event_header header;
-	__u32 pid, tid;
-	__u64 start;
-	__u64 len;
-	__u64 pgoff;
+	u32 pid, tid;
+	u64 start;
+	u64 len;
+	u64 pgoff;
 	char filename[PATH_MAX];
 };
 
 struct comm_event {
 	struct perf_event_header header;
-	__u32 pid, tid;
+	u32 pid, tid;
 	char comm[16];
 };
 
 struct fork_event {
 	struct perf_event_header header;
-	__u32 pid, ppid;
+	u32 pid, ppid;
 };
 
 struct period_event {
 	struct perf_event_header header;
-	__u64 time;
-	__u64 id;
-	__u64 sample_period;
+	u64 time;
+	u64 id;
+	u64 sample_period;
 };
 
 typedef union event_union {
@@ -84,6 +90,13 @@ typedef union event_union {
 	struct period_event		period;
 } event_t;
 
+
+struct sym_ext {
+	struct rb_node	node;
+	double		percent;
+	char		*path;
+};
+
 static LIST_HEAD(dsos);
 static struct dso *kernel_dso;
 static struct dso *vdso;
@@ -145,7 +158,7 @@ static void dsos__fprintf(FILE *fp)
 		dso__fprintf(pos, fp);
 }
 
-static struct symbol *vdso__find_symbol(struct dso *dso, __u64 ip)
+static struct symbol *vdso__find_symbol(struct dso *dso, u64 ip)
 {
 	return dso__find_symbol(kernel_dso, ip);
 }
@@ -178,19 +191,19 @@ static int load_kernel(void)
 
 struct map {
 	struct list_head node;
-	__u64	 start;
-	__u64	 end;
-	__u64	 pgoff;
-	__u64	 (*map_ip)(struct map *, __u64);
+	u64	 start;
+	u64	 end;
+	u64	 pgoff;
+	u64	 (*map_ip)(struct map *, u64);
 	struct dso	 *dso;
 };
 
-static __u64 map__map_ip(struct map *map, __u64 ip)
+static u64 map__map_ip(struct map *map, u64 ip)
 {
 	return ip - map->start + map->pgoff;
 }
 
-static __u64 vdso__map_ip(struct map *map, __u64 ip)
+static u64 vdso__map_ip(struct map *map, u64 ip)
 {
 	return ip;
 }
@@ -373,7 +386,7 @@ static int thread__fork(struct thread *self, struct thread *parent)
 	return 0;
 }
 
-static struct map *thread__find_map(struct thread *self, __u64 ip)
+static struct map *thread__find_map(struct thread *self, u64 ip)
 {
 	struct map *pos;
 
@@ -414,7 +427,7 @@ struct hist_entry {
 	struct map	 *map;
 	struct dso	 *dso;
 	struct symbol	 *sym;
-	__u64	 ip;
+	u64	 ip;
 	char		 level;
 
 	uint32_t	 count;
@@ -519,7 +532,7 @@ sort__dso_print(FILE *fp, struct hist_entry *self)
 	if (self->dso)
 		return fprintf(fp, "%-25s", self->dso->name);
 
-	return fprintf(fp, "%016llx         ", (__u64)self->ip);
+	return fprintf(fp, "%016llx         ", (u64)self->ip);
 }
 
 static struct sort_entry sort_dso = {
@@ -533,7 +546,7 @@ static struct sort_entry sort_dso = {
 static int64_t
 sort__sym_cmp(struct hist_entry *left, struct hist_entry *right)
 {
-	__u64 ip_l, ip_r;
+	u64 ip_l, ip_r;
 
 	if (left->sym == right->sym)
 		return 0;
@@ -550,13 +563,13 @@ sort__sym_print(FILE *fp, struct hist_entry *self)
 	size_t ret = 0;
 
 	if (verbose)
-		ret += fprintf(fp, "%#018llx  ", (__u64)self->ip);
+		ret += fprintf(fp, "%#018llx  ", (u64)self->ip);
 
 	if (self->sym) {
 		ret += fprintf(fp, "[%c] %s",
 			self->dso == kernel_dso ? 'k' : '.', self->sym->name);
 	} else {
-		ret += fprintf(fp, "%#016llx", (__u64)self->ip);
+		ret += fprintf(fp, "%#016llx", (u64)self->ip);
 	}
 
 	return ret;
@@ -647,7 +660,7 @@ hist_entry__collapse(struct hist_entry *left, struct hist_entry *right)
 /*
  * collect histogram counts
  */
-static void hist_hit(struct hist_entry *he, __u64 ip)
+static void hist_hit(struct hist_entry *he, u64 ip)
 {
 	unsigned int sym_size, offset;
 	struct symbol *sym = he->sym;
@@ -676,7 +689,7 @@ static void hist_hit(struct hist_entry *he, __u64 ip)
 
 static int
 hist_entry__add(struct thread *thread, struct map *map, struct dso *dso,
-		struct symbol *sym, __u64 ip, char level)
+		struct symbol *sym, u64 ip, char level)
 {
 	struct rb_node **p = &hist.rb_node;
 	struct rb_node *parent = NULL;
@@ -848,7 +861,7 @@ process_overflow_event(event_t *event, unsigned long offset, unsigned long head)
 	int show = 0;
 	struct dso *dso = NULL;
 	struct thread *thread = threads__findnew(event->ip.pid);
-	__u64 ip = event->ip.ip;
+	u64 ip = event->ip.ip;
 	struct map *map = NULL;
 
 	dprintf("%p [%p]: PERF_EVENT (IP, %d): %d: %p\n",
@@ -1030,13 +1043,33 @@ process_event(event_t *event, unsigned long offset, unsigned long head)
 	return 0;
 }
 
+static char *get_color(double percent)
+{
+	char *color = PERF_COLOR_NORMAL;
+
+	/*
+	 * We color high-overhead entries in red, mid-overhead
+	 * entries in green - and keep the low overhead places
+	 * normal:
+	 */
+	if (percent >= MIN_RED)
+		color = PERF_COLOR_RED;
+	else {
+		if (percent > MIN_GREEN)
+			color = PERF_COLOR_GREEN;
+	}
+	return color;
+}
+
 static int
-parse_line(FILE *file, struct symbol *sym, __u64 start, __u64 len)
+parse_line(FILE *file, struct symbol *sym, u64 start, u64 len)
 {
 	char *line = NULL, *tmp, *tmp2;
+	static const char *prev_line;
+	static const char *prev_color;
 	unsigned int offset;
 	size_t line_len;
-	__u64 line_ip;
+	u64 line_ip;
 	int ret;
 	char *c;
 
@@ -1073,27 +1106,36 @@ parse_line(FILE *file, struct symbol *sym, __u64 start, __u64 len)
 	}
 
 	if (line_ip != -1) {
+		const char *path = NULL;
 		unsigned int hits = 0;
 		double percent = 0.0;
-		char *color = PERF_COLOR_NORMAL;
+		char *color;
+		struct sym_ext *sym_ext = sym->priv;
 
 		offset = line_ip - start;
 		if (offset < len)
 			hits = sym->hist[offset];
 
-		if (sym->hist_sum)
+		if (offset < len && sym_ext) {
+			path = sym_ext[offset].path;
+			percent = sym_ext[offset].percent;
+		} else if (sym->hist_sum)
 			percent = 100.0 * hits / sym->hist_sum;
 
+		color = get_color(percent);
+
 		/*
-		 * We color high-overhead entries in red, mid-overhead
-		 * entries in green - and keep the low overhead places
-		 * normal:
+		 * Also color the filename and line if needed, with
+		 * the same color than the percentage. Don't print it
+		 * twice for close colored ip with the same filename:line
 		 */
-		if (percent >= 5.0)
-			color = PERF_COLOR_RED;
-		else {
-			if (percent > 0.5)
-				color = PERF_COLOR_GREEN;
+		if (path) {
+			if (!prev_line || strcmp(prev_line, path)
+				       || color != prev_color) {
+				color_fprintf(stdout, color, " %s", path);
+				prev_line = path;
+				prev_color = color;
+			}
 		}
 
 		color_fprintf(stdout, color, " %7.2f", percent);
@@ -1109,10 +1151,125 @@ parse_line(FILE *file, struct symbol *sym, __u64 start, __u64 len)
 	return 0;
 }
 
+static struct rb_root root_sym_ext;
+
+static void insert_source_line(struct sym_ext *sym_ext)
+{
+	struct sym_ext *iter;
+	struct rb_node **p = &root_sym_ext.rb_node;
+	struct rb_node *parent = NULL;
+
+	while (*p != NULL) {
+		parent = *p;
+		iter = rb_entry(parent, struct sym_ext, node);
+
+		if (sym_ext->percent > iter->percent)
+			p = &(*p)->rb_left;
+		else
+			p = &(*p)->rb_right;
+	}
+
+	rb_link_node(&sym_ext->node, parent, p);
+	rb_insert_color(&sym_ext->node, &root_sym_ext);
+}
+
+static void free_source_line(struct symbol *sym, int len)
+{
+	struct sym_ext *sym_ext = sym->priv;
+	int i;
+
+	if (!sym_ext)
+		return;
+
+	for (i = 0; i < len; i++)
+		free(sym_ext[i].path);
+	free(sym_ext);
+
+	sym->priv = NULL;
+	root_sym_ext = RB_ROOT;
+}
+
+/* Get the filename:line for the colored entries */
+static void
+get_source_line(struct symbol *sym, u64 start, int len, char *filename)
+{
+	int i;
+	char cmd[PATH_MAX * 2];
+	struct sym_ext *sym_ext;
+
+	if (!sym->hist_sum)
+		return;
+
+	sym->priv = calloc(len, sizeof(struct sym_ext));
+	if (!sym->priv)
+		return;
+
+	sym_ext = sym->priv;
+
+	for (i = 0; i < len; i++) {
+		char *path = NULL;
+		size_t line_len;
+		u64 offset;
+		FILE *fp;
+
+		sym_ext[i].percent = 100.0 * sym->hist[i] / sym->hist_sum;
+		if (sym_ext[i].percent <= 0.5)
+			continue;
+
+		offset = start + i;
+		sprintf(cmd, "addr2line -e %s %016llx", filename, offset);
+		fp = popen(cmd, "r");
+		if (!fp)
+			continue;
+
+		if (getline(&path, &line_len, fp) < 0 || !line_len)
+			goto next;
+
+		sym_ext[i].path = malloc(sizeof(char) * line_len + 1);
+		if (!sym_ext[i].path)
+			goto next;
+
+		strcpy(sym_ext[i].path, path);
+		insert_source_line(&sym_ext[i]);
+
+	next:
+		pclose(fp);
+	}
+}
+
+static void print_summary(char *filename)
+{
+	struct sym_ext *sym_ext;
+	struct rb_node *node;
+
+	printf("\nSorted summary for file %s\n", filename);
+	printf("----------------------------------------------\n\n");
+
+	if (RB_EMPTY_ROOT(&root_sym_ext)) {
+		printf(" Nothing higher than %1.1f%%\n", MIN_GREEN);
+		return;
+	}
+
+	node = rb_first(&root_sym_ext);
+	while (node) {
+		double percent;
+		char *color;
+		char *path;
+
+		sym_ext = rb_entry(node, struct sym_ext, node);
+		percent = sym_ext->percent;
+		color = get_color(percent);
+		path = sym_ext->path;
+
+		color_fprintf(stdout, color, " %7.2f %s", percent, path);
+		node = rb_next(node);
+	}
+}
+
 static void annotate_sym(struct dso *dso, struct symbol *sym)
 {
 	char *filename = dso->name;
-	__u64 start, end, len;
+	u64 start, end, len;
 	char command[PATH_MAX*2];
 	FILE *file;
 
@@ -1121,13 +1278,6 @@ static void annotate_sym(struct dso *dso, struct symbol *sym)
 	if (dso == kernel_dso)
 		filename = vmlinux;
 
-	printf("\n------------------------------------------------\n");
-	printf(" Percent |	Source code & Disassembly of %s\n", filename);
-	printf("------------------------------------------------\n");
-
-	if (verbose >= 2)
-		printf("annotating [%p] %30s : [%p] %30s\n", dso, dso->name, sym, sym->name);
-
 	start = sym->obj_start;
 	if (!start)
 		start = sym->start;
@@ -1135,7 +1285,19 @@ static void annotate_sym(struct dso *dso, struct symbol *sym)
 	end = start + sym->end - sym->start + 1;
 	len = sym->end - sym->start;
 
-	sprintf(command, "objdump --start-address=0x%016Lx --stop-address=0x%016Lx -dS %s", (__u64)start, (__u64)end, filename);
+	if (print_line) {
+		get_source_line(sym, start, len, filename);
+		print_summary(filename);
+	}
+
+	printf("\n\n------------------------------------------------\n");
+	printf(" Percent |	Source code & Disassembly of %s\n", filename);
+	printf("------------------------------------------------\n");
+
+	if (verbose >= 2)
+		printf("annotating [%p] %30s : [%p] %30s\n", dso, dso->name, sym, sym->name);
+
+	sprintf(command, "objdump --start-address=0x%016Lx --stop-address=0x%016Lx -dS %s", (u64)start, (u64)end, filename);
 
 	if (verbose >= 3)
 		printf("doing: %s\n", command);
@@ -1150,6 +1312,8 @@ static void annotate_sym(struct dso *dso, struct symbol *sym)
 	}
 
 	pclose(file);
+	if (print_line)
+		free_source_line(sym, len);
 }
 
 static void find_annotations(void)
@@ -1308,6 +1472,8 @@ static const struct option options[] = {
 	OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace,
 		    "dump raw trace in ASCII"),
 	OPT_STRING('k', "vmlinux", &vmlinux, "file", "vmlinux pathname"),
+	OPT_BOOLEAN('l', "print-line", &print_line,
+		    "print matching source lines (may be slow)"),
 	OPT_END()
 };
 
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 0f5771f615d..d7ebbd75754 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -37,33 +37,37 @@ static pid_t			target_pid			= -1;
 static int			inherit				= 1;
 static int			force				= 0;
 static int			append_file			= 0;
+static int			call_graph			= 0;
 static int			verbose				= 0;
 
 static long			samples;
 static struct timeval		last_read;
 static struct timeval		this_read;
 
-static __u64			bytes_written;
+static u64			bytes_written;
 
 static struct pollfd		event_array[MAX_NR_CPUS * MAX_COUNTERS];
 
 static int			nr_poll;
 static int			nr_cpu;
 
+static int			file_new = 1;
+static struct perf_file_header	file_header;
+
 struct mmap_event {
 	struct perf_event_header	header;
-	__u32				pid;
-	__u32				tid;
-	__u64				start;
-	__u64				len;
-	__u64				pgoff;
+	u32				pid;
+	u32				tid;
+	u64				start;
+	u64				len;
+	u64				pgoff;
 	char				filename[PATH_MAX];
 };
 
 struct comm_event {
 	struct perf_event_header	header;
-	__u32				pid;
-	__u32				tid;
+	u32				pid;
+	u32				tid;
 	char				comm[16];
 };
 
@@ -77,10 +81,10 @@ struct mmap_data {
 
 static struct mmap_data		mmap_array[MAX_NR_CPUS][MAX_COUNTERS];
 
-static unsigned int mmap_read_head(struct mmap_data *md)
+static unsigned long mmap_read_head(struct mmap_data *md)
 {
 	struct perf_counter_mmap_page *pc = md->base;
-	int head;
+	long head;
 
 	head = pc->data_head;
 	rmb();
@@ -88,6 +92,32 @@ static unsigned int mmap_read_head(struct mmap_data *md)
 	return head;
 }
 
+static void mmap_write_tail(struct mmap_data *md, unsigned long tail)
+{
+	struct perf_counter_mmap_page *pc = md->base;
+
+	/*
+	 * ensure all reads are done before we write the tail out.
+	 */
+	/* mb(); */
+	pc->data_tail = tail;
+}
+
+static void write_output(void *buf, size_t size)
+{
+	while (size) {
+		int ret = write(output, buf, size);
+
+		if (ret < 0)
+			die("failed to write");
+
+		size -= ret;
+		buf += ret;
+
+		bytes_written += ret;
+	}
+}
+
 static void mmap_read(struct mmap_data *md)
 {
 	unsigned int head = mmap_read_head(md);
@@ -108,7 +138,7 @@ static void mmap_read(struct mmap_data *md)
 	 * In either case, truncate and restart at head.
 	 */
 	diff = head - old;
-	if (diff > md->mask / 2 || diff < 0) {
+	if (diff < 0) {
 		struct timeval iv;
 		unsigned long msecs;
 
@@ -136,36 +166,17 @@ static void mmap_read(struct mmap_data *md)
 		size = md->mask + 1 - (old & md->mask);
 		old += size;
 
-		while (size) {
-			int ret = write(output, buf, size);
-
-			if (ret < 0)
-				die("failed to write");
-
-			size -= ret;
-			buf += ret;
-
-			bytes_written += ret;
-		}
+		write_output(buf, size);
 	}
 
 	buf = &data[old & md->mask];
 	size = head - old;
 	old += size;
 
-	while (size) {
-		int ret = write(output, buf, size);
-
-		if (ret < 0)
-			die("failed to write");
-
-		size -= ret;
-		buf += ret;
-
-		bytes_written += ret;
-	}
+	write_output(buf, size);
 
 	md->prev = old;
+	mmap_write_tail(md, old);
 }
 
 static volatile int done = 0;
@@ -191,7 +202,7 @@ static void pid_synthesize_comm_event(pid_t pid, int full)
 	struct comm_event comm_ev;
 	char filename[PATH_MAX];
 	char bf[BUFSIZ];
-	int fd, ret;
+	int fd;
 	size_t size;
 	char *field, *sep;
 	DIR *tasks;
@@ -201,8 +212,12 @@ static void pid_synthesize_comm_event(pid_t pid, int full)
 
 	fd = open(filename, O_RDONLY);
 	if (fd < 0) {
-		fprintf(stderr, "couldn't open %s\n", filename);
-		exit(EXIT_FAILURE);
+		/*
+		 * We raced with a task exiting - just return:
+		 */
+		if (verbose)
+			fprintf(stderr, "couldn't open %s\n", filename);
+		return;
 	}
 	if (read(fd, bf, sizeof(bf)) < 0) {
 		fprintf(stderr, "couldn't read %s\n", filename);
@@ -223,17 +238,13 @@ static void pid_synthesize_comm_event(pid_t pid, int full)
 
 	comm_ev.pid = pid;
 	comm_ev.header.type = PERF_EVENT_COMM;
-	size = ALIGN(size, sizeof(__u64));
+	size = ALIGN(size, sizeof(u64));
 	comm_ev.header.size = sizeof(comm_ev) - (sizeof(comm_ev.comm) - size);
 
 	if (!full) {
 		comm_ev.tid = pid;
 
-		ret = write(output, &comm_ev, comm_ev.header.size);
-		if (ret < 0) {
-			perror("failed to write");
-			exit(-1);
-		}
+		write_output(&comm_ev, comm_ev.header.size);
 		return;
 	}
 
@@ -248,11 +259,7 @@ static void pid_synthesize_comm_event(pid_t pid, int full)
 
 		comm_ev.tid = pid;
 
-		ret = write(output, &comm_ev, comm_ev.header.size);
-		if (ret < 0) {
-			perror("failed to write");
-			exit(-1);
-		}
+		write_output(&comm_ev, comm_ev.header.size);
 	}
 	closedir(tasks);
 	return;
@@ -272,8 +279,12 @@ static void pid_synthesize_mmap_samples(pid_t pid)
 
 	fp = fopen(filename, "r");
 	if (fp == NULL) {
-		fprintf(stderr, "couldn't open %s\n", filename);
-		exit(EXIT_FAILURE);
+		/*
+		 * We raced with a task exiting - just return:
+		 */
+		if (verbose)
+			fprintf(stderr, "couldn't open %s\n", filename);
+		return;
 	}
 	while (1) {
 		char bf[BUFSIZ], *pbf = bf;
@@ -304,17 +315,14 @@ static void pid_synthesize_mmap_samples(pid_t pid)
 			size = strlen(execname);
 			execname[size - 1] = '\0'; /* Remove \n */
 			memcpy(mmap_ev.filename, execname, size);
-			size = ALIGN(size, sizeof(__u64));
+			size = ALIGN(size, sizeof(u64));
 			mmap_ev.len -= mmap_ev.start;
 			mmap_ev.header.size = (sizeof(mmap_ev) -
 					       (sizeof(mmap_ev.filename) - size));
 			mmap_ev.pid = pid;
 			mmap_ev.tid = pid;
 
-			if (write(output, &mmap_ev, mmap_ev.header.size) < 0) {
-				perror("failed to write");
-				exit(-1);
-			}
+			write_output(&mmap_ev, mmap_ev.header.size);
 		}
 	}
 
@@ -351,11 +359,25 @@ static void create_counter(int counter, int cpu, pid_t pid)
 	int track = 1;
 
 	attr->sample_type	= PERF_SAMPLE_IP | PERF_SAMPLE_TID;
+
 	if (freq) {
 		attr->sample_type	|= PERF_SAMPLE_PERIOD;
 		attr->freq		= 1;
 		attr->sample_freq	= freq;
 	}
+
+	if (call_graph)
+		attr->sample_type	|= PERF_SAMPLE_CALLCHAIN;
+
+	if (file_new) {
+		file_header.sample_type = attr->sample_type;
+	} else {
+		if (file_header.sample_type != attr->sample_type) {
+			fprintf(stderr, "incompatible append\n");
+			exit(-1);
+		}
+	}
+
 	attr->mmap		= track;
 	attr->comm		= track;
 	attr->inherit		= (cpu < 0) && inherit;
@@ -410,7 +432,7 @@ try_again:
 	mmap_array[nr_cpu][counter].prev = 0;
 	mmap_array[nr_cpu][counter].mask = mmap_pages*page_size - 1;
 	mmap_array[nr_cpu][counter].base = mmap(NULL, (mmap_pages+1)*page_size,
-			PROT_READ, MAP_SHARED, fd[nr_cpu][counter], 0);
+			PROT_READ|PROT_WRITE, MAP_SHARED, fd[nr_cpu][counter], 0);
 	if (mmap_array[nr_cpu][counter].base == MAP_FAILED) {
 		error("failed to mmap with %d (%s)\n", errno, strerror(errno));
 		exit(-1);
@@ -435,6 +457,14 @@ static void open_counters(int cpu, pid_t pid)
 	nr_cpu++;
 }
 
+static void atexit_header(void)
+{
+	file_header.data_size += bytes_written;
+
+	if (pwrite(output, &file_header, sizeof(file_header), 0) == -1)
+		perror("failed to write on file headers");
+}
+
 static int __cmd_record(int argc, const char **argv)
 {
 	int i, counter;
@@ -448,6 +478,10 @@ static int __cmd_record(int argc, const char **argv)
 	assert(nr_cpus <= MAX_NR_CPUS);
 	assert(nr_cpus >= 0);
 
+	atexit(sig_atexit);
+	signal(SIGCHLD, sig_handler);
+	signal(SIGINT, sig_handler);
+
 	if (!stat(output_name, &st) && !force && !append_file) {
 		fprintf(stderr, "Error, output file %s exists, use -A to append or -f to overwrite.\n",
 				output_name);
@@ -456,7 +490,7 @@ static int __cmd_record(int argc, const char **argv)
 
 	flags = O_CREAT|O_RDWR;
 	if (append_file)
-		flags |= O_APPEND;
+		file_new = 0;
 	else
 		flags |= O_TRUNC;
 
@@ -466,15 +500,22 @@ static int __cmd_record(int argc, const char **argv)
 		exit(-1);
 	}
 
+	if (!file_new) {
+		if (read(output, &file_header, sizeof(file_header)) == -1) {
+			perror("failed to read file headers");
+			exit(-1);
+		}
+
+		lseek(output, file_header.data_size, SEEK_CUR);
+	}
+
+	atexit(atexit_header);
+
 	if (!system_wide) {
 		open_counters(-1, target_pid != -1 ? target_pid : getpid());
 	} else for (i = 0; i < nr_cpus; i++)
 		open_counters(i, target_pid);
 
-	atexit(sig_atexit);
-	signal(SIGCHLD, sig_handler);
-	signal(SIGINT, sig_handler);
-
 	if (target_pid == -1 && argc) {
 		pid = fork();
 		if (pid < 0)
@@ -555,6 +596,8 @@ static const struct option options[] = {
 		    "profile at this frequency"),
 	OPT_INTEGER('m', "mmap-pages", &mmap_pages,
 		    "number of mmap data pages"),
+	OPT_BOOLEAN('g', "call-graph", &call_graph,
+		    "do call-graph (stack chain/backtrace) recording"),
 	OPT_BOOLEAN('v', "verbose", &verbose,
 		    "be more verbose (show counter open errors, etc)"),
 	OPT_END()
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index 82fa93b4db9..5eb5566f0c9 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -36,45 +36,65 @@ static int		show_mask = SHOW_KERNEL | SHOW_USER | SHOW_HV;
 
 static int		dump_trace = 0;
 #define dprintf(x...)	do { if (dump_trace) printf(x); } while (0)
+#define cdprintf(x...)	do { if (dump_trace) color_fprintf(stdout, color, x); } while (0)
 
 static int		verbose;
+#define eprintf(x...)	do { if (verbose) fprintf(stderr, x); } while (0)
+
 static int		full_paths;
 
 static unsigned long	page_size;
 static unsigned long	mmap_window = 32;
 
+static char		default_parent_pattern[] = "^sys_|^do_page_fault";
+static char		*parent_pattern = default_parent_pattern;
+static regex_t		parent_regex;
+
+static int		exclude_other = 1;
+
 struct ip_event {
 	struct perf_event_header header;
-	__u64 ip;
-	__u32 pid, tid;
-	__u64 period;
+	u64 ip;
+	u32 pid, tid;
+	unsigned char __more_data[];
+};
+
+struct ip_callchain {
+	u64 nr;
+	u64 ips[0];
 };
 
 struct mmap_event {
 	struct perf_event_header header;
-	__u32 pid, tid;
-	__u64 start;
-	__u64 len;
-	__u64 pgoff;
+	u32 pid, tid;
+	u64 start;
+	u64 len;
+	u64 pgoff;
 	char filename[PATH_MAX];
 };
 
 struct comm_event {
 	struct perf_event_header header;
-	__u32 pid, tid;
+	u32 pid, tid;
 	char comm[16];
 };
 
 struct fork_event {
 	struct perf_event_header header;
-	__u32 pid, ppid;
+	u32 pid, ppid;
 };
 
 struct period_event {
 	struct perf_event_header header;
-	__u64 time;
-	__u64 id;
-	__u64 sample_period;
+	u64 time;
+	u64 id;
+	u64 sample_period;
+};
+
+struct lost_event {
+	struct perf_event_header header;
+	u64 id;
+	u64 lost;
 };
 
 typedef union event_union {
@@ -84,6 +104,7 @@ typedef union event_union {
 	struct comm_event		comm;
 	struct fork_event		fork;
 	struct period_event		period;
+	struct lost_event		lost;
 } event_t;
 
 static LIST_HEAD(dsos);
@@ -119,15 +140,11 @@ static struct dso *dsos__findnew(const char *name)
 
 	nr = dso__load(dso, NULL, verbose);
 	if (nr < 0) {
-		if (verbose)
-			fprintf(stderr, "Failed to open: %s\n", name);
+		eprintf("Failed to open: %s\n", name);
 		goto out_delete_dso;
 	}
-	if (!nr && verbose) {
-		fprintf(stderr,
-		"No symbols found in: %s, maybe install a debug package?\n",
-				name);
-	}
+	if (!nr)
+		eprintf("No symbols found in: %s, maybe install a debug package?\n", name);
 
 	dsos__add(dso);
 
@@ -146,7 +163,7 @@ static void dsos__fprintf(FILE *fp)
 		dso__fprintf(pos, fp);
 }
 
-static struct symbol *vdso__find_symbol(struct dso *dso, __u64 ip)
+static struct symbol *vdso__find_symbol(struct dso *dso, u64 ip)
 {
 	return dso__find_symbol(kernel_dso, ip);
 }
@@ -193,19 +210,19 @@ static int strcommon(const char *pathname)
 
 struct map {
 	struct list_head node;
-	__u64	 start;
-	__u64	 end;
-	__u64	 pgoff;
-	__u64	 (*map_ip)(struct map *, __u64);
+	u64	 start;
+	u64	 end;
+	u64	 pgoff;
+	u64	 (*map_ip)(struct map *, u64);
 	struct dso	 *dso;
 };
 
-static __u64 map__map_ip(struct map *map, __u64 ip)
+static u64 map__map_ip(struct map *map, u64 ip)
 {
 	return ip - map->start + map->pgoff;
 }
 
-static __u64 vdso__map_ip(struct map *map, __u64 ip)
+static u64 vdso__map_ip(struct map *map, u64 ip)
 {
 	return ip;
 }
@@ -412,7 +429,7 @@ static int thread__fork(struct thread *self, struct thread *parent)
 	return 0;
 }
 
-static struct map *thread__find_map(struct thread *self, __u64 ip)
+static struct map *thread__find_map(struct thread *self, u64 ip)
 {
 	struct map *pos;
 
@@ -453,10 +470,11 @@ struct hist_entry {
 	struct map	 *map;
 	struct dso	 *dso;
 	struct symbol	 *sym;
-	__u64		 ip;
+	struct symbol	 *parent;
+	u64		 ip;
 	char		 level;
 
-	__u64		 count;
+	u64		 count;
 };
 
 /*
@@ -473,6 +491,16 @@ struct sort_entry {
 	size_t	(*print)(FILE *fp, struct hist_entry *);
 };
 
+static int64_t cmp_null(void *l, void *r)
+{
+	if (!l && !r)
+		return 0;
+	else if (!l)
+		return -1;
+	else
+		return 1;
+}
+
 /* --sort pid */
 
 static int64_t
@@ -507,14 +535,8 @@ sort__comm_collapse(struct hist_entry *left, struct hist_entry *right)
 	char *comm_l = left->thread->comm;
 	char *comm_r = right->thread->comm;
 
-	if (!comm_l || !comm_r) {
-		if (!comm_l && !comm_r)
-			return 0;
-		else if (!comm_l)
-			return -1;
-		else
-			return 1;
-	}
+	if (!comm_l || !comm_r)
+		return cmp_null(comm_l, comm_r);
 
 	return strcmp(comm_l, comm_r);
 }
@@ -540,14 +562,8 @@ sort__dso_cmp(struct hist_entry *left, struct hist_entry *right)
 	struct dso *dso_l = left->dso;
 	struct dso *dso_r = right->dso;
 
-	if (!dso_l || !dso_r) {
-		if (!dso_l && !dso_r)
-			return 0;
-		else if (!dso_l)
-			return -1;
-		else
-			return 1;
-	}
+	if (!dso_l || !dso_r)
+		return cmp_null(dso_l, dso_r);
 
 	return strcmp(dso_l->name, dso_r->name);
 }
@@ -558,7 +574,7 @@ sort__dso_print(FILE *fp, struct hist_entry *self)
 	if (self->dso)
 		return fprintf(fp, "%-25s", self->dso->name);
 
-	return fprintf(fp, "%016llx         ", (__u64)self->ip);
+	return fprintf(fp, "%016llx         ", (u64)self->ip);
 }
 
 static struct sort_entry sort_dso = {
@@ -572,7 +588,7 @@ static struct sort_entry sort_dso = {
 static int64_t
 sort__sym_cmp(struct hist_entry *left, struct hist_entry *right)
 {
-	__u64 ip_l, ip_r;
+	u64 ip_l, ip_r;
 
 	if (left->sym == right->sym)
 		return 0;
@@ -589,13 +605,13 @@ sort__sym_print(FILE *fp, struct hist_entry *self)
 	size_t ret = 0;
 
 	if (verbose)
-		ret += fprintf(fp, "%#018llx  ", (__u64)self->ip);
+		ret += fprintf(fp, "%#018llx  ", (u64)self->ip);
 
 	if (self->sym) {
 		ret += fprintf(fp, "[%c] %s",
 			self->dso == kernel_dso ? 'k' : '.', self->sym->name);
 	} else {
-		ret += fprintf(fp, "%#016llx", (__u64)self->ip);
+		ret += fprintf(fp, "%#016llx", (u64)self->ip);
 	}
 
 	return ret;
@@ -607,7 +623,38 @@ static struct sort_entry sort_sym = {
 	.print	= sort__sym_print,
 };
 
+/* --sort parent */
+
+static int64_t
+sort__parent_cmp(struct hist_entry *left, struct hist_entry *right)
+{
+	struct symbol *sym_l = left->parent;
+	struct symbol *sym_r = right->parent;
+
+	if (!sym_l || !sym_r)
+		return cmp_null(sym_l, sym_r);
+
+	return strcmp(sym_l->name, sym_r->name);
+}
+
+static size_t
+sort__parent_print(FILE *fp, struct hist_entry *self)
+{
+	size_t ret = 0;
+
+	ret += fprintf(fp, "%-20s", self->parent ? self->parent->name : "[other]");
+
+	return ret;
+}
+
+static struct sort_entry sort_parent = {
+	.header = "Parent symbol       ",
+	.cmp	= sort__parent_cmp,
+	.print	= sort__parent_print,
+};
+
 static int sort__need_collapse = 0;
+static int sort__has_parent = 0;
 
 struct sort_dimension {
 	char			*name;
@@ -620,6 +667,7 @@ static struct sort_dimension sort_dimensions[] = {
 	{ .name = "comm",	.entry = &sort_comm,	},
 	{ .name = "dso",	.entry = &sort_dso,	},
 	{ .name = "symbol",	.entry = &sort_sym,	},
+	{ .name = "parent",	.entry = &sort_parent,	},
 };
 
 static LIST_HEAD(hist_entry__sort_list);
@@ -640,6 +688,19 @@ static int sort_dimension__add(char *tok)
 		if (sd->entry->collapse)
 			sort__need_collapse = 1;
 
+		if (sd->entry == &sort_pa