[PATCH] perf tools: Unify perf.data mapping and events handling

[Frederic Weisbecker]

This librarizes the perf.data file mapping and handling in various
perf tools, roughly reducing the amount of code and fixing the places
that mmap from beginning of the file whereas we want to mmap from
the beginning of the data, leading to page faults because the mmap
window is too small since the trace info are written in the file too.

TODO:

- Convert perf timechart too
- Fix strange huge number of total samples reported in perf report

Signed-off-by: Frederic Weisbecker <fweisbec@xxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Arnaldo Carvalho de Melo <acme@xxxxxxxxxx>
Cc: Mike Galbraith <efault@xxxxxx>
Cc: Paul Mackerras <paulus@xxxxxxxxx>
---
 tools/perf/Makefile         |    2 +
 tools/perf/builtin-report.c |  211 +++++++------------------------------------
 tools/perf/builtin-sched.c  |  140 ++++++-----------------------
 tools/perf/builtin-trace.c  |  129 ++++-----------------------
 4 files changed, 81 insertions(+), 401 deletions(-)

diff --git a/tools/perf/Makefile b/tools/perf/Makefile
index 5a42996..495eb6d 100644
--- a/tools/perf/Makefile
+++ b/tools/perf/Makefile
@@ -342,6 +342,7 @@ LIB_H += util/values.h
 LIB_H += util/sort.h
 LIB_H += util/hist.h
 LIB_H += util/thread.h
+LIB_H += util/data_map.h
 
 LIB_OBJS += util/abspath.o
 LIB_OBJS += util/alias.o
@@ -378,6 +379,7 @@ LIB_OBJS += util/trace-event-info.o
 LIB_OBJS += util/svghelper.o
 LIB_OBJS += util/sort.o
 LIB_OBJS += util/hist.o
+LIB_OBJS += util/data_map.o
 
 BUILTIN_OBJS += builtin-annotate.o
 BUILTIN_OBJS += builtin-help.o
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index 12f8c86..87c4582 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -26,6 +26,7 @@
 #include "util/parse-options.h"
 #include "util/parse-events.h"
 
+#include "util/data_map.h"
 #include "util/thread.h"
 #include "util/sort.h"
 #include "util/hist.h"
@@ -37,7 +38,6 @@ static char		*dso_list_str, *comm_list_str, *sym_list_str,
 static struct strlist	*dso_list, *comm_list, *sym_list;
 
 static int		force;
-static int		input;
 
 static int		full_paths;
 static int		show_nr_samples;
@@ -48,15 +48,11 @@ static struct perf_read_values	show_threads_values;
 static char		default_pretty_printing_style[] = "normal";
 static char		*pretty_printing_style = default_pretty_printing_style;
 
-static unsigned long	page_size;
-static unsigned long	mmap_window = 32;
-
 static int		exclude_other = 1;
 
 static char		callchain_default_opt[] = "fractal,0.5";
 
-static char		__cwd[PATH_MAX];
-static char		*cwd = __cwd;
+static char		*cwd;
 static int		cwdlen;
 
 static struct rb_root	threads;
@@ -815,208 +811,71 @@ process_read_event(event_t *event, unsigned long offset, unsigned long head)
 	return 0;
 }
 
-static int
-process_event(event_t *event, unsigned long offset, unsigned long head)
-{
-	trace_event(event);
-
-	switch (event->header.type) {
-	case PERF_RECORD_SAMPLE:
-		return process_sample_event(event, offset, head);
-
-	case PERF_RECORD_MMAP:
-		return process_mmap_event(event, offset, head);
-
-	case PERF_RECORD_COMM:
-		return process_comm_event(event, offset, head);
-
-	case PERF_RECORD_FORK:
-	case PERF_RECORD_EXIT:
-		return process_task_event(event, offset, head);
-
-	case PERF_RECORD_LOST:
-		return process_lost_event(event, offset, head);
-
-	case PERF_RECORD_READ:
-		return process_read_event(event, offset, head);
-
-	/*
-	 * We dont process them right now but they are fine:
-	 */
-
-	case PERF_RECORD_THROTTLE:
-	case PERF_RECORD_UNTHROTTLE:
-		return 0;
-
-	default:
-		return -1;
-	}
-
-	return 0;
-}
-
-static int __cmd_report(void)
+static int sample_type_check(u64 type)
 {
-	int ret, rc = EXIT_FAILURE;
-	unsigned long offset = 0;
-	unsigned long head, shift;
-	struct stat input_stat;
-	struct thread *idle;
-	event_t *event;
-	uint32_t size;
-	char *buf;
-
-	idle = register_idle_thread(&threads, &last_match);
-	thread__comm_adjust(idle);
-
-	if (show_threads)
-		perf_read_values_init(&show_threads_values);
-
-	input = open(input_name, O_RDONLY);
-	if (input < 0) {
-		fprintf(stderr, " failed to open file: %s", input_name);
-		if (!strcmp(input_name, "perf.data"))
-			fprintf(stderr, "  (try 'perf record' first)");
-		fprintf(stderr, "\n");
-		exit(-1);
-	}
-
-	ret = fstat(input, &input_stat);
-	if (ret < 0) {
-		perror("failed to stat file");
-		exit(-1);
-	}
-
-	if (!force && input_stat.st_uid && (input_stat.st_uid != geteuid())) {
-		fprintf(stderr, "file: %s not owned by current user or root\n", input_name);
-		exit(-1);
-	}
-
-	if (!input_stat.st_size) {
-		fprintf(stderr, "zero-sized file, nothing to do!\n");
-		exit(0);
-	}
-
-	header = perf_header__read(input);
-	head = header->data_offset;
-
-	sample_type = perf_header__sample_type(header);
+	sample_type = type;
 
 	if (!(sample_type & PERF_SAMPLE_CALLCHAIN)) {
 		if (sort__has_parent) {
 			fprintf(stderr, "selected --sort parent, but no"
 					" callchain data. Did you call"
 					" perf record without -g?\n");
-			exit(-1);
+			return -1;
 		}
 		if (callchain) {
 			fprintf(stderr, "selected -g but no callchain data."
 					" Did you call perf record without"
 					" -g?\n");
-			exit(-1);
+			return -1;
 		}
 	} else if (callchain_param.mode != CHAIN_NONE && !callchain) {
 			callchain = 1;
 			if (register_callchain_param(&callchain_param) < 0) {
 				fprintf(stderr, "Can't register callchain"
 						" params\n");
-				exit(-1);
+				return -1;
 			}
 	}
 
-	if (load_kernel() < 0) {
-		perror("failed to load kernel symbols");
-		return EXIT_FAILURE;
-	}
-
-	if (!full_paths) {
-		if (getcwd(__cwd, sizeof(__cwd)) == NULL) {
-			perror("failed to get the current directory");
-			return EXIT_FAILURE;
-		}
-		cwdlen = strlen(cwd);
-	} else {
-		cwd = NULL;
-		cwdlen = 0;
-	}
-
-	shift = page_size * (head / page_size);
-	offset += shift;
-	head -= shift;
-
-remap:
-	buf = (char *)mmap(NULL, page_size * mmap_window, PROT_READ,
-			   MAP_SHARED, input, offset);
-	if (buf == MAP_FAILED) {
-		perror("failed to mmap file");
-		exit(-1);
-	}
-
-more:
-	event = (event_t *)(buf + head);
-
-	size = event->header.size;
-	if (!size)
-		size = 8;
-
-	if (head + event->header.size >= page_size * mmap_window) {
-		int munmap_ret;
-
-		shift = page_size * (head / page_size);
-
-		munmap_ret = munmap(buf, page_size * mmap_window);
-		assert(munmap_ret == 0);
-
-		offset += shift;
-		head -= shift;
-		goto remap;
-	}
-
-	size = event->header.size;
-
-	dump_printf("\n%p [%p]: event: %d\n",
-			(void *)(offset + head),
-			(void *)(long)event->header.size,
-			event->header.type);
-
-	if (!size || process_event(event, offset, head) < 0) {
-
-		dump_printf("%p [%p]: skipping unknown header type: %d\n",
-			(void *)(offset + head),
-			(void *)(long)(event->header.size),
-			event->header.type);
-
-		total_unknown++;
+	return 0;
+}
 
-		/*
-		 * assume we lost track of the stream, check alignment, and
-		 * increment a single u64 in the hope to catch on again 'soon'.
-		 */
+static struct perf_file_handler file_handler = {
+	.process_sample_event	= process_sample_event,
+	.process_mmap_event	= process_mmap_event,
+	.process_comm_event	= process_comm_event,
+	.process_exit_event	= process_task_event,
+	.process_fork_event	= process_task_event,
+	.process_lost_event	= process_lost_event,
+	.process_read_event	= process_read_event,
+	.sample_type_check	= sample_type_check,
+};
 
-		if (unlikely(head & 7))
-			head &= ~7ULL;
 
-		size = 8;
-	}
+static int __cmd_report(void)
+{
+	struct thread *idle;
+	int ret;
 
-	head += size;
+	idle = register_idle_thread(&threads, &last_match);
+	thread__comm_adjust(idle);
 
-	if (offset + head >= header->data_offset + header->data_size)
-		goto done;
+	if (show_threads)
+		perf_read_values_init(&show_threads_values);
 
-	if (offset + head < (unsigned long)input_stat.st_size)
-		goto more;
+	register_perf_file_handler(&file_handler);
 
-done:
-	rc = EXIT_SUCCESS;
-	close(input);
+	ret = mmap_dispatch_perf_file(&header, input_name, force, full_paths,
+				      &cwdlen, &cwd);
+	if (ret)
+		return ret;
 
 	dump_printf("      IP events: %10ld\n", total);
 	dump_printf("    mmap events: %10ld\n", total_mmap);
 	dump_printf("    comm events: %10ld\n", total_comm);
 	dump_printf("    fork events: %10ld\n", total_fork);
 	dump_printf("    lost events: %10ld\n", total_lost);
-	dump_printf(" unknown events: %10ld\n", total_unknown);
+	dump_printf(" unknown events: %10ld\n", file_handler.total_unknown);
 
 	if (dump_trace)
 		return 0;
@@ -1034,7 +893,7 @@ done:
 	if (show_threads)
 		perf_read_values_destroy(&show_threads_values);
 
-	return rc;
+	return ret;
 }
 
 static int
@@ -1177,8 +1036,6 @@ int cmd_report(int argc, const char **argv, const char *prefix __used)
 {
 	symbol__init();
 
-	page_size = getpagesize();
-
 	argc = parse_options(argc, argv, options, report_usage, 0);
 
 	setup_sorting();
diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
index 1887138..e1df705 100644
--- a/tools/perf/builtin-sched.c
+++ b/tools/perf/builtin-sched.c
@@ -11,6 +11,7 @@
 #include "util/trace-event.h"
 
 #include "util/debug.h"
+#include "util/data_map.h"
 
 #include <sys/types.h>
 #include <sys/prctl.h>
@@ -20,9 +21,6 @@
 #include <math.h>
 
 static char			const *input_name = "perf.data";
-static int			input;
-static unsigned long		page_size;
-static unsigned long		mmap_window = 32;
 
 static unsigned long		total_comm = 0;
 
@@ -35,6 +33,9 @@ static u64			sample_type;
 static char			default_sort_order[] = "avg, max, switch, runtime";
 static char			*sort_order = default_sort_order;
 
+static char			*cwd;
+static int			cwdlen;
+
 #define PR_SET_NAME		15               /* Set process name */
 #define MAX_CPUS		4096
 
@@ -1594,129 +1595,43 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head)
 }
 
 static int
-process_event(event_t *event, unsigned long offset, unsigned long head)
+process_lost_event(event_t *event __used,
+		   unsigned long offset __used,
+		   unsigned long head __used)
 {
-	trace_event(event);
-
-	nr_events++;
-	switch (event->header.type) {
-	case PERF_RECORD_MMAP:
-		return 0;
-	case PERF_RECORD_LOST:
-		nr_lost_chunks++;
-		nr_lost_events += event->lost.lost;
-		return 0;
-
-	case PERF_RECORD_COMM:
-		return process_comm_event(event, offset, head);
+	nr_lost_chunks++;
+	nr_lost_events += event->lost.lost;
 
-	case PERF_RECORD_EXIT ... PERF_RECORD_READ:
-		return 0;
+	return 0;
+}
 
-	case PERF_RECORD_SAMPLE:
-		return process_sample_event(event, offset, head);
+static int sample_type_check(u64 type)
+{
+	sample_type = type;
 
-	case PERF_RECORD_MAX:
-	default:
+	if (!(sample_type & PERF_SAMPLE_RAW)) {
+		fprintf(stderr,
+			"No trace sample to read. Did you call perf record "
+			"without -R?");
 		return -1;
 	}
 
 	return 0;
 }
 
+static struct perf_file_handler file_handler = {
+	.process_sample_event	= process_sample_event,
+	.process_comm_event	= process_comm_event,
+	.process_lost_event	= process_lost_event,
+	.sample_type_check	= sample_type_check,
+};
+
 static int read_events(void)
 {
-	int ret, rc = EXIT_FAILURE;
-	unsigned long offset = 0;
-	unsigned long head = 0;
-	struct stat perf_stat;
-	event_t *event;
-	uint32_t size;
-	char *buf;
-
 	register_idle_thread(&threads, &last_match);
+	register_perf_file_handler(&file_handler);
 
-	input = open(input_name, O_RDONLY);
-	if (input < 0) {
-		perror("failed to open file");
-		exit(-1);
-	}
-
-	ret = fstat(input, &perf_stat);
-	if (ret < 0) {
-		perror("failed to stat file");
-		exit(-1);
-	}
-
-	if (!perf_stat.st_size) {
-		fprintf(stderr, "zero-sized file, nothing to do!\n");
-		exit(0);
-	}
-	header = perf_header__read(input);
-	head = header->data_offset;
-	sample_type = perf_header__sample_type(header);
-
-	if (!(sample_type & PERF_SAMPLE_RAW))
-		die("No trace sample to read. Did you call perf record "
-		    "without -R?");
-
-	if (load_kernel() < 0) {
-		perror("failed to load kernel symbols");
-		return EXIT_FAILURE;
-	}
-
-remap:
-	buf = (char *)mmap(NULL, page_size * mmap_window, PROT_READ,
-			   MAP_SHARED, input, offset);
-	if (buf == MAP_FAILED) {
-		perror("failed to mmap file");
-		exit(-1);
-	}
-
-more:
-	event = (event_t *)(buf + head);
-
-	size = event->header.size;
-	if (!size)
-		size = 8;
-
-	if (head + event->header.size >= page_size * mmap_window) {
-		unsigned long shift = page_size * (head / page_size);
-		int res;
-
-		res = munmap(buf, page_size * mmap_window);
-		assert(res == 0);
-
-		offset += shift;
-		head -= shift;
-		goto remap;
-	}
-
-	size = event->header.size;
-
-
-	if (!size || process_event(event, offset, head) < 0) {
-
-		/*
-		 * assume we lost track of the stream, check alignment, and
-		 * increment a single u64 in the hope to catch on again 'soon'.
-		 */
-
-		if (unlikely(head & 7))
-			head &= ~7ULL;
-
-		size = 8;
-	}
-
-	head += size;
-
-	if (offset + head < (unsigned long)perf_stat.st_size)
-		goto more;
-
-	rc = EXIT_SUCCESS;
-	close(input);
-
-	return rc;
+	return mmap_dispatch_perf_file(&header, input_name, 0, 0, &cwdlen, &cwd);
 }
 
 static void print_bad_events(void)
@@ -1934,7 +1849,6 @@ static int __cmd_record(int argc, const char **argv)
 int cmd_sched(int argc, const char **argv, const char *prefix __used)
 {
 	symbol__init();
-	page_size = getpagesize();
 
 	argc = parse_options(argc, argv, sched_options, sched_usage,
 			     PARSE_OPT_STOP_AT_NON_OPTION);
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index d9abb4a..fb3f3c2 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -12,11 +12,9 @@
 #include "util/debug.h"
 
 #include "util/trace-event.h"
+#include "util/data_map.h"
 
 static char		const *input_name = "perf.data";
-static int		input;
-static unsigned long	page_size;
-static unsigned long	mmap_window = 32;
 
 static unsigned long	total = 0;
 static unsigned long	total_comm = 0;
@@ -27,6 +25,9 @@ static struct thread	*last_match;
 static struct perf_header *header;
 static u64		sample_type;
 
+static char		*cwd;
+static int		cwdlen;
+
 
 static int
 process_comm_event(event_t *event, unsigned long offset, unsigned long head)
@@ -112,125 +113,32 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head)
 	return 0;
 }
 
-static int
-process_event(event_t *event, unsigned long offset, unsigned long head)
+static int sample_type_check(u64 type)
 {
-	trace_event(event);
-
-	switch (event->header.type) {
-	case PERF_RECORD_MMAP ... PERF_RECORD_LOST:
-		return 0;
-
-	case PERF_RECORD_COMM:
-		return process_comm_event(event, offset, head);
-
-	case PERF_RECORD_EXIT ... PERF_RECORD_READ:
-		return 0;
+	sample_type = type;
 
-	case PERF_RECORD_SAMPLE:
-		return process_sample_event(event, offset, head);
-
-	case PERF_RECORD_MAX:
-	default:
+	if (!(sample_type & PERF_SAMPLE_RAW)) {
+		fprintf(stderr,
+			"No trace sample to read. Did you call perf record "
+			"without -R?");
 		return -1;
 	}
 
 	return 0;
 }
 
+static struct perf_file_handler file_handler = {
+	.process_sample_event	= process_sample_event,
+	.process_comm_event	= process_comm_event,
+	.sample_type_check	= sample_type_check,
+};
+
 static int __cmd_trace(void)
 {
-	int ret, rc = EXIT_FAILURE;
-	unsigned long offset = 0;
-	unsigned long head = 0;
-	unsigned long shift;
-	struct stat perf_stat;
-	event_t *event;
-	uint32_t size;
-	char *buf;
-
 	register_idle_thread(&threads, &last_match);
+	register_perf_file_handler(&file_handler);
 
-	input = open(input_name, O_RDONLY);
-	if (input < 0) {
-		perror("failed to open file");
-		exit(-1);
-	}
-
-	ret = fstat(input, &perf_stat);
-	if (ret < 0) {
-		perror("failed to stat file");
-		exit(-1);
-	}
-
-	if (!perf_stat.st_size) {
-		fprintf(stderr, "zero-sized file, nothing to do!\n");
-		exit(0);
-	}
-	header = perf_header__read(input);
-	head = header->data_offset;
-	sample_type = perf_header__sample_type(header);
-
-	if (!(sample_type & PERF_SAMPLE_RAW))
-		die("No trace sample to read. Did you call perf record "
-		    "without -R?");
-
-	if (load_kernel() < 0) {
-		perror("failed to load kernel symbols");
-		return EXIT_FAILURE;
-	}
-
-	shift = page_size * (head / page_size);
-	offset += shift;
-	head -= shift;
-
-remap:
-	buf = (char *)mmap(NULL, page_size * mmap_window, PROT_READ,
-			   MAP_SHARED, input, offset);
-	if (buf == MAP_FAILED) {
-		perror("failed to mmap file");
-		exit(-1);
-	}
-
-more:
-	event = (event_t *)(buf + head);
-
-	if (head + event->header.size >= page_size * mmap_window) {
-		int res;
-
-		shift = page_size * (head / page_size);
-		res = munmap(buf, page_size * mmap_window);
-		assert(res == 0);
-
-		offset += shift;
-		head -= shift;
-		goto remap;
-	}
-
-	size = event->header.size;
-
-	if (!size || process_event(event, offset, head) < 0) {
-
-		/*
-		 * assume we lost track of the stream, check alignment, and
-		 * increment a single u64 in the hope to catch on again 'soon'.
-		 */
-
-		if (unlikely(head & 7))
-			head &= ~7ULL;
-
-		size = 8;
-	}
-
-	head += size;
-
-	if (offset + head < (unsigned long)perf_stat.st_size)
-		goto more;
-
-	rc = EXIT_SUCCESS;
-	close(input);
-
-	return rc;
+	return mmap_dispatch_perf_file(&header, input_name, 0, 0, &cwdlen, &cwd);
 }
 
 static const char * const annotate_usage[] = {
@@ -249,7 +157,6 @@ static const struct option options[] = {
 int cmd_trace(int argc, const char **argv, const char *prefix __used)
 {
 	symbol__init();
-	page_size = getpagesize();
 
 	argc = parse_options(argc, argv, options, annotate_usage, 0);
 	if (argc) {
-- 
1.6.2.3

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/