tools: bpf: move the event reading loop to libbpf

There are two copies of event reading loop - in bpftool and
trace_helpers "library".  Consolidate them and move the code
to libbpf.  Return codes from trace_helpers are kept, but
renamed to include LIBBPF prefix.

Suggested-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

samples/bpf/Makefile

@@ -165,6 +165,14 @@ HOSTCFLAGS += -I$(srctree)/tools/lib/ -I$(srctree)/tools/include
 HOSTCFLAGS += -I$(srctree)/tools/perf
 
 HOSTCFLAGS_bpf_load.o += -I$(objtree)/usr/include -Wno-unused-variable
+HOSTCFLAGS_trace_helpers.o += -I$(srctree)/tools/lib/bpf/
+
+HOSTCFLAGS_trace_output_user.o += -I$(srctree)/tools/lib/bpf/
+HOSTCFLAGS_offwaketime_user.o += -I$(srctree)/tools/lib/bpf/
+HOSTCFLAGS_spintest_user.o += -I$(srctree)/tools/lib/bpf/
+HOSTCFLAGS_trace_event_user.o += -I$(srctree)/tools/lib/bpf/
+HOSTCFLAGS_sampleip_user.o += -I$(srctree)/tools/lib/bpf/
+
 HOSTLOADLIBES_test_lru_dist += -lelf
 HOSTLOADLIBES_sock_example += -lelf
 HOSTLOADLIBES_fds_example += -lelf

samples/bpf/trace_output_user.c

@@ -48,7 +48,7 @@ static int print_bpf_output(void *data, int size)
 	if (e->cookie != 0x12345678) {
 		printf("BUG pid %llx cookie %llx sized %d\n",
 		       e->pid, e->cookie, size);
-		return PERF_EVENT_ERROR;
+		return LIBBPF_PERF_EVENT_ERROR;
 	}
 
 	cnt++;
@@ -56,10 +56,10 @@ static int print_bpf_output(void *data, int size)
 	if (cnt == MAX_CNT) {
 		printf("recv %lld events per sec\n",
 		       MAX_CNT * 1000000000ll / (time_get_ns() - start_time));
-		return PERF_EVENT_DONE;
+		return LIBBPF_PERF_EVENT_DONE;
 	}
 
-	return PERF_EVENT_CONT;
+	return LIBBPF_PERF_EVENT_CONT;
 }
 
 static void test_bpf_perf_event(void)

tools/bpf/bpftool/map_perf_ring.c

@@ -50,14 +50,15 @@ static void int_exit(int signo)
 	stop = true;
 }
 
-static void
-print_bpf_output(struct event_ring_info *ring, struct perf_event_sample *e)
+static enum bpf_perf_event_ret print_bpf_output(void *event, void *priv)
 {
+	struct event_ring_info *ring = priv;
+	struct perf_event_sample *e = event;
 	struct {
 		struct perf_event_header header;
 		__u64 id;
 		__u64 lost;
-	} *lost = (void *)e;
+	} *lost = event;
 
 	if (json_output) {
 		jsonw_start_object(json_wtr);
@@ -96,60 +97,23 @@ print_bpf_output(struct event_ring_info *ring, struct perf_event_sample *e)
 			       e->header.type, e->header.size);
 		}
 	}
+
+	return LIBBPF_PERF_EVENT_CONT;
 }
 
 static void
 perf_event_read(struct event_ring_info *ring, void **buf, size_t *buf_len)
 {
-	volatile struct perf_event_mmap_page *header = ring->mem;
-	__u64 buffer_size = MMAP_PAGE_CNT * get_page_size();
-	__u64 data_tail = header->data_tail;
-	__u64 data_head = header->data_head;
-	void *base, *begin, *end;
+	enum bpf_perf_event_ret ret;
 
-	asm volatile("" ::: "memory"); /* in real code it should be smp_rmb() */
-	if (data_head == data_tail)
-		return;
-
-	base = ((char *)header) + get_page_size();
-
-	begin = base + data_tail % buffer_size;
-	end = base + data_head % buffer_size;
-
-	while (begin != end) {
-		struct perf_event_sample *e;
-
-		e = begin;
-		if (begin + e->header.size > base + buffer_size) {
-			long len = base + buffer_size - begin;
-
-			if (*buf_len < e->header.size) {
-				free(*buf);
-				*buf = malloc(e->header.size);
-				if (!*buf) {
-					fprintf(stderr,
-						"can't allocate memory");
-					stop = true;
-					return;
-				}
-				*buf_len = e->header.size;
-			}
-
-			memcpy(*buf, begin, len);
-			memcpy(*buf + len, base, e->header.size - len);
-			e = (void *)*buf;
-			begin = base + e->header.size - len;
-		} else if (begin + e->header.size == base + buffer_size) {
-			begin = base;
-		} else {
-			begin += e->header.size;
-		}
-
-		print_bpf_output(ring, e);
+	ret = bpf_perf_event_read_simple(ring->mem,
+					 MMAP_PAGE_CNT * get_page_size(),
+					 get_page_size(), buf, buf_len,
+					 print_bpf_output, ring);
+	if (ret != LIBBPF_PERF_EVENT_CONT) {
+		fprintf(stderr, "perf read loop failed with %d\n", ret);
+		stop = true;
 	}
-
-	__sync_synchronize(); /* smp_mb() */
-	header->data_tail = data_head;
 }
 
 static int perf_mmap_size(void)

tools/lib/bpf/Makefile

@@ -69,7 +69,7 @@ FEATURE_USER = .libbpf
 FEATURE_TESTS = libelf libelf-getphdrnum libelf-mmap bpf
 FEATURE_DISPLAY = libelf bpf
 
-INCLUDES = -I. -I$(srctree)/tools/include -I$(srctree)/tools/arch/$(ARCH)/include/uapi -I$(srctree)/tools/include/uapi
+INCLUDES = -I. -I$(srctree)/tools/include -I$(srctree)/tools/arch/$(ARCH)/include/uapi -I$(srctree)/tools/include/uapi -I$(srctree)/tools/perf
 FEATURE_CHECK_CFLAGS-bpf = $(INCLUDES)
 
 check_feat := 1

tools/lib/bpf/libbpf.c

@@ -31,6 +31,7 @@
 #include <unistd.h>
 #include <fcntl.h>
 #include <errno.h>
+#include <perf-sys.h>
 #include <asm/unistd.h>
 #include <linux/err.h>
 #include <linux/kernel.h>
@@ -2210,3 +2211,63 @@ int bpf_prog_load_xattr(const struct bpf_prog_load_attr *attr,
 	*prog_fd = bpf_program__fd(first_prog);
 	return 0;
 }
+
+enum bpf_perf_event_ret
+bpf_perf_event_read_simple(void *mem, unsigned long size,
+			   unsigned long page_size, void **buf, size_t *buf_len,
+			   bpf_perf_event_print_t fn, void *priv)
+{
+	volatile struct perf_event_mmap_page *header = mem;
+	__u64 data_tail = header->data_tail;
+	__u64 data_head = header->data_head;
+	void *base, *begin, *end;
+	int ret;
+
+	asm volatile("" ::: "memory"); /* in real code it should be smp_rmb() */
+	if (data_head == data_tail)
+		return LIBBPF_PERF_EVENT_CONT;
+
+	base = ((char *)header) + page_size;
+
+	begin = base + data_tail % size;
+	end = base + data_head % size;
+
+	while (begin != end) {
+		struct perf_event_header *ehdr;
+
+		ehdr = begin;
+		if (begin + ehdr->size > base + size) {
+			long len = base + size - begin;
+
+			if (*buf_len < ehdr->size) {
+				free(*buf);
+				*buf = malloc(ehdr->size);
+				if (!*buf) {
+					ret = LIBBPF_PERF_EVENT_ERROR;
+					break;
+				}
+				*buf_len = ehdr->size;
+			}
+
+			memcpy(*buf, begin, len);
+			memcpy(*buf + len, base, ehdr->size - len);
+			ehdr = (void *)*buf;
+			begin = base + ehdr->size - len;
+		} else if (begin + ehdr->size == base + size) {
+			begin = base;
+		} else {
+			begin += ehdr->size;
+		}
+
+		ret = fn(ehdr, priv);
+		if (ret != LIBBPF_PERF_EVENT_CONT)
+			break;
+
+		data_tail += ehdr->size;
+	}
+
+	__sync_synchronize(); /* smp_mb() */
+	header->data_tail = data_tail;
+
+	return ret;
+}

tools/lib/bpf/libbpf.h

@@ -267,4 +267,17 @@ int bpf_prog_load(const char *file, enum bpf_prog_type type,
 		  struct bpf_object **pobj, int *prog_fd);
 
 int bpf_set_link_xdp_fd(int ifindex, int fd, __u32 flags);
+
+enum bpf_perf_event_ret {
+	LIBBPF_PERF_EVENT_DONE	= 0,
+	LIBBPF_PERF_EVENT_ERROR	= -1,
+	LIBBPF_PERF_EVENT_CONT	= -2,
+};
+
+typedef enum bpf_perf_event_ret (*bpf_perf_event_print_t)(void *event,
+							  void *priv);
+int bpf_perf_event_read_simple(void *mem, unsigned long size,
+			       unsigned long page_size,
+			       void **buf, size_t *buf_len,
+			       bpf_perf_event_print_t fn, void *priv);
 #endif

tools/testing/selftests/bpf/Makefile

@@ -10,7 +10,7 @@ ifneq ($(wildcard $(GENHDR)),)
   GENFLAGS := -DHAVE_GENHDR
 endif
 
-CFLAGS += -Wall -O2 -I$(APIDIR) -I$(LIBDIR) -I$(GENDIR) $(GENFLAGS) -I../../../include
+CFLAGS += -Wall -O2 -I$(APIDIR) -I$(LIBDIR) -I$(BPFDIR) -I$(GENDIR) $(GENFLAGS) -I../../../include
 LDLIBS += -lcap -lelf -lrt -lpthread
 
 TEST_CUSTOM_PROGS = $(OUTPUT)/urandom_read

tools/testing/selftests/bpf/test_progs.c

@@ -1337,12 +1337,12 @@ static int get_stack_print_output(void *data, int size)
 			good_user_stack = true;
 	}
 	if (!good_kern_stack || !good_user_stack)
-		return PERF_EVENT_ERROR;
+		return LIBBPF_PERF_EVENT_ERROR;
 
 	if (cnt == MAX_CNT_RAWTP)
-		return PERF_EVENT_DONE;
+		return LIBBPF_PERF_EVENT_DONE;
 
-	return PERF_EVENT_CONT;
+	return LIBBPF_PERF_EVENT_CONT;
 }
 
 static void test_get_stack_raw_tp(void)

tools/testing/selftests/bpf/trace_helpers.c

@@ -74,7 +74,7 @@ struct ksym *ksym_search(long key)
 
 static int page_size;
 static int page_cnt = 8;
-static volatile struct perf_event_mmap_page *header;
+static struct perf_event_mmap_page *header;
 
 int perf_event_mmap(int fd)
 {
@@ -107,74 +107,47 @@ struct perf_event_sample {
 	char data[];
 };
 
-static int perf_event_read(perf_event_print_fn fn)
+static enum bpf_perf_event_ret bpf_perf_event_print(void *event, void *priv)
 {
-	__u64 data_tail = header->data_tail;
-	__u64 data_head = header->data_head;
-	__u64 buffer_size = page_cnt * page_size;
-	void *base, *begin, *end;
-	char buf[256];
+	struct perf_event_sample *e = event;
+	perf_event_print_fn fn = priv;
 	int ret;
 
-	asm volatile("" ::: "memory"); /* in real code it should be smp_rmb() */
-	if (data_head == data_tail)
-		return PERF_EVENT_CONT;
-
-	base = ((char *)header) + page_size;
-
-	begin = base + data_tail % buffer_size;
-	end = base + data_head % buffer_size;
-
-	while (begin != end) {
-		struct perf_event_sample *e;
-
-		e = begin;
-		if (begin + e->header.size > base + buffer_size) {
-			long len = base + buffer_size - begin;
-
-			assert(len < e->header.size);
-			memcpy(buf, begin, len);
-			memcpy(buf + len, base, e->header.size - len);
-			e = (void *) buf;
-			begin = base + e->header.size - len;
-		} else if (begin + e->header.size == base + buffer_size) {
-			begin = base;
-		} else {
-			begin += e->header.size;
-		}
-
-		if (e->header.type == PERF_RECORD_SAMPLE) {
-			ret = fn(e->data, e->size);
-			if (ret != PERF_EVENT_CONT)
-				return ret;
-		} else if (e->header.type == PERF_RECORD_LOST) {
-			struct {
-				struct perf_event_header header;
-				__u64 id;
-				__u64 lost;
-			} *lost = (void *) e;
-			printf("lost %lld events\n", lost->lost);
-		} else {
-			printf("unknown event type=%d size=%d\n",
-			       e->header.type, e->header.size);
-		}
+	if (e->header.type == PERF_RECORD_SAMPLE) {
+		ret = fn(e->data, e->size);
+		if (ret != LIBBPF_PERF_EVENT_CONT)
+			return ret;
+	} else if (e->header.type == PERF_RECORD_LOST) {
+		struct {
+			struct perf_event_header header;
+			__u64 id;
+			__u64 lost;
+		} *lost = (void *) e;
+		printf("lost %lld events\n", lost->lost);
+	} else {
+		printf("unknown event type=%d size=%d\n",
+		       e->header.type, e->header.size);
 	}
 
-	__sync_synchronize(); /* smp_mb() */
-	header->data_tail = data_head;
-	return PERF_EVENT_CONT;
+	return LIBBPF_PERF_EVENT_CONT;
 }
 
 int perf_event_poller(int fd, perf_event_print_fn output_fn)
 {
-	int ret;
+	enum bpf_perf_event_ret ret;
+	void *buf = NULL;
+	size_t len = 0;
 
 	for (;;) {
 		perf_event_poll(fd);
-		ret = perf_event_read(output_fn);
-		if (ret != PERF_EVENT_CONT)
-			return ret;
+		ret = bpf_perf_event_read_simple(header, page_cnt * page_size,
+						 page_size, &buf, &len,
+						 bpf_perf_event_print,
+						 output_fn);
+		if (ret != LIBBPF_PERF_EVENT_CONT)
+			break;
 	}
+	free(buf);
 
-	return PERF_EVENT_DONE;
+	return ret;
 }

tools/testing/selftests/bpf/trace_helpers.h

@@ -2,6 +2,8 @@
 #ifndef __TRACE_HELPER_H
 #define __TRACE_HELPER_H
 
+#include <libbpf.h>
+
 struct ksym {
 	long addr;
 	char *name;
@@ -10,14 +12,9 @@ struct ksym {
 int load_kallsyms(void);
 struct ksym *ksym_search(long key);
 
-typedef int (*perf_event_print_fn)(void *data, int size);
-
-/* return code for perf_event_print_fn */
-#define PERF_EVENT_DONE		0
-#define PERF_EVENT_ERROR	-1
-#define PERF_EVENT_CONT		-2
+typedef enum bpf_perf_event_ret (*perf_event_print_fn)(void *data, int size);
 
 int perf_event_mmap(int fd);
-/* return PERF_EVENT_DONE or PERF_EVENT_ERROR */
+/* return LIBBPF_PERF_EVENT_DONE or LIBBPF_PERF_EVENT_ERROR */
 int perf_event_poller(int fd, perf_event_print_fn output_fn);
 #endif