[RFC 2/2] samples: bpf: IO profiling example

[He Kuang]

For the first node, we sample and record its characteristic
(eg. address), and add the result to a global hash table.  For other
nodes, they'll be traced when their context matches the characteristics
in previous tables.

Signed-off-by: He Kuang <hekuang@xxxxxxxxxx>
---
 samples/bpf/Makefile       |   4 +
 samples/bpf/tracex5_kern.c | 195 +++++++++++++++++++++++++++++++++++++++++++++
 samples/bpf/tracex5_user.c |  56 +++++++++++++
 3 files changed, 255 insertions(+)
 create mode 100644 samples/bpf/tracex5_kern.c
 create mode 100644 samples/bpf/tracex5_user.c

diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile
index aec1f52..529deb7 100644
--- a/samples/bpf/Makefile
+++ b/samples/bpf/Makefile
@@ -11,6 +11,7 @@ hostprogs-y += tracex1
 hostprogs-y += tracex2
 hostprogs-y += tracex3
 hostprogs-y += tracex4
+hostprogs-y += tracex5
 
 dropmon-objs := dropmon.o libbpf.o
 test_verifier-objs := test_verifier.o libbpf.o
@@ -22,6 +23,7 @@ tracex1-objs := bpf_load.o libbpf.o tracex1_user.o
 tracex2-objs := bpf_load.o libbpf.o tracex2_user.o
 tracex3-objs := bpf_load.o libbpf.o tracex3_user.o
 tracex4-objs := bpf_load.o libbpf.o tracex4_user.o
+tracex5-objs := bpf_load.o libbpf.o tracex5_user.o
 
 # Tell kbuild to always build the programs
 always := $(hostprogs-y)
@@ -31,6 +33,7 @@ always += tracex1_kern.o
 always += tracex2_kern.o
 always += tracex3_kern.o
 always += tracex4_kern.o
+always += tracex5_kern.o
 
 HOSTCFLAGS += -I$(objtree)/usr/include
 HOSTCFLAGS += -I$(objtree)/include/uapi/
@@ -43,6 +46,7 @@ HOSTLOADLIBES_tracex1 += -lelf
 HOSTLOADLIBES_tracex2 += -lelf
 HOSTLOADLIBES_tracex3 += -lelf
 HOSTLOADLIBES_tracex4 += -lelf
+HOSTLOADLIBES_tracex5 += -lelf
 
 # point this to your LLVM backend with bpf support
 #LLC=$(srctree)/tools/bpf/llvm/bld/Debug+Asserts/bin/llc
diff --git a/samples/bpf/tracex5_kern.c b/samples/bpf/tracex5_kern.c
new file mode 100644
index 0000000..67c03b0
--- /dev/null
+++ b/samples/bpf/tracex5_kern.c
@@ -0,0 +1,195 @@
+/*
+  In this example, we placed several nodes in kernel source, all these
+  nodes in the data stream path, and represent the data transfer.
+
+  we sampling on the first node and record the characteristic in it, add
+  to a global hash table, then the next node will be traced when it's
+  context matches the characteristics in previous tables.'
+*/
+#include <linux/skbuff.h>
+#include <linux/netdevice.h>
+#include <uapi/linux/bpf.h>
+#include <trace/bpf_trace.h>
+#include "bpf_helpers.h"
+
+#include <linux/writeback.h>
+#include <linux/blkdev.h>
+
+struct globals {
+	int num_samples;
+	int samples;
+};
+
+struct bpf_map_def SEC("maps") global_map = {
+	.type = BPF_MAP_TYPE_ARRAY,
+	.key_size = sizeof(int),
+	.value_size = sizeof(struct globals),
+	.max_entries = 1,
+};
+
+/* trigger table: symbol(page) */
+struct bpf_map_def SEC("maps") trigger_page_hash = {
+	.type = BPF_MAP_TYPE_HASH,
+	.key_size = sizeof(u64),
+	.value_size = sizeof(int),
+	.max_entries = 512,
+};
+
+/* trigger table: symbol(bio) */
+struct bpf_map_def SEC("maps") trigger_bio_hash = {
+	.type = BPF_MAP_TYPE_HASH,
+	.key_size = sizeof(u64),
+	.value_size = sizeof(int),
+	.max_entries = 512,
+};
+
+
+/* trigger table symbol(req) */
+struct bpf_map_def SEC("maps") trigger_req_hash = {
+	.type = BPF_MAP_TYPE_HASH,
+	.key_size = sizeof(u64),
+	.value_size = sizeof(int),
+	.max_entries = 512,
+};
+
+//iov -> page
+SEC("events/filemap/iov_iter_copy_from_user_atomic")
+int node1(struct bpf_context *ctx)
+{
+	int ind = 0;
+	struct globals *g = bpf_map_lookup_elem(&global_map, &ind);
+
+	if (!g)
+		return 0;
+
+	g->num_samples++;
+	if ((g->num_samples >> 10) * 1024 == g->num_samples) {
+		struct page* page = (struct page*)ctx->arg1;
+		int fill = 1;
+
+		__sync_fetch_and_add(&g->samples, 1);
+
+		bpf_map_update_elem(&trigger_page_hash, &page, &fill, BPF_ANY);
+
+		return 1;
+	} else {
+		return 0;
+	}
+}
+
+/*
+  ext4.h is not in the include pathes, it's not convenient to fetch
+  struct fields now.
+
+  Though in this example we should avoid this, but if we want to use
+  bio->bi_iter, we should use the struct definition.
+*/
+struct ext4_io_submit {
+	int			io_op;
+	struct bio		*io_bio;
+};
+
+//page -> bio
+SEC("events/ext4/ext4_bio_write_page")
+int node_2(struct bpf_context *ctx)
+{
+	u64 page_addr = ctx->arg2;
+	struct ext4_io_submit* io = (struct ext4_io_submit*)ctx->arg1;
+	struct bio* bio = bpf_fetch_ptr(&io->io_bio);
+	void *value;
+	int fill = 1;
+
+	value = bpf_map_lookup_elem(&trigger_page_hash, &page_addr);
+	if (!value)
+		return 0;
+
+	bpf_map_update_elem(&trigger_bio_hash, &bio, &fill, BPF_ANY);
+	return 1;
+}
+
+//bio -> rq case1
+SEC("events/block/blk_queue_bio")
+int node_3_1(struct bpf_context *ctx)
+{
+	struct request *request = (struct request*)ctx->arg1;
+	struct bio* bio = (struct bio * )ctx->arg2;
+	void *value;
+	int fill = 1;
+	int i;
+	int found = 0;
+
+	value = bpf_map_lookup_elem(&trigger_bio_hash, &bio);
+	if (!value) {
+		return 0;
+	}
+
+	bpf_map_update_elem(&trigger_req_hash, &request, &fill, BPF_ANY);
+	return 1;
+}
+
+//bio -> rq case2
+SEC("events/block/bio_attempt_front_merge")
+int node_3_2(struct bpf_context *ctx)
+{
+	struct request *request = (struct request*)ctx->arg1;
+	struct bio* bio = (struct bio * )ctx->arg2;
+	void *value;
+	int fill = 1;
+
+	value = bpf_map_lookup_elem(&trigger_bio_hash, &bio);
+	if (!value) {
+		return 0;
+	}
+
+	bpf_map_update_elem(&trigger_req_hash, &request, &fill, BPF_ANY);
+	return 1;
+}
+
+//bio -> rq case3
+SEC("events/block/bio_attempt_back_merge")
+int node_3_3(struct bpf_context *ctx)
+{
+	struct request *request = (struct request*)ctx->arg1;
+	struct bio* bio = (struct bio * )ctx->arg2;
+	void *value;
+	int fill = 1;
+
+	value = bpf_map_lookup_elem(&trigger_bio_hash, &bio);
+	if (!value) {
+		return 0;
+	}
+
+	bpf_map_update_elem(&trigger_req_hash, &request, &fill, BPF_ANY);
+	return 1;
+}
+
+//req -> vq
+SEC("events/scsi/virtscsi_add_cmd")
+int node_4(struct bpf_context *ctx)
+{
+	struct request *request = (struct request*)ctx->arg2;
+	void *value;
+	int fill = 1;
+
+	value = bpf_map_lookup_elem(&trigger_req_hash, &request);
+	if (!value) {
+		return 0;
+	}
+	return 1;
+}
+
+//end: bio
+SEC("events/block/bio_endio")
+int node_5(struct bpf_context *ctx)
+{
+	struct bio* bio = (struct bio*)ctx->arg1;
+	void *value;
+
+	value = bpf_map_lookup_elem(&trigger_bio_hash, &bio);
+	if (!value) {
+		return 0;
+	}
+	return 1;
+}
+
+char _license[] SEC("license") = "GPL";
diff --git a/samples/bpf/tracex5_user.c b/samples/bpf/tracex5_user.c
new file mode 100644
index 0000000..814f926
--- /dev/null
+++ b/samples/bpf/tracex5_user.c
@@ -0,0 +1,56 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <unistd.h>
+
+#include <linux/bpf.h>
+#include "libbpf.h"
+#include "bpf_load.h"
+
+struct globals {
+	int num_samples;
+	int samples;
+};
+
+static void print_hist(int fd)
+{
+	struct globals g = {};
+	int key = 0;
+
+	bpf_lookup_elem(map_fd[0], &key, &g);
+
+	printf("Total samples = %d, num=%d\n", g.samples, g.num_samples);
+}
+
+static void int_exit(int sig)
+{
+	print_hist(map_fd[0]);
+	exit(0);
+}
+
+int main(int ac, char **argv)
+{
+	FILE *f;
+	char filename[256];
+
+	snprintf(filename, sizeof(filename), "%s_kern.o", argv[0]);
+
+	if (load_bpf_file(filename)) {
+		printf("%s", bpf_log_buf);
+		return 1;
+	}
+
+	/* start 'dd' in the background to have plenty of 'write' syscalls */
+	f = popen("dd if=/dev/zero of=/mnt/data/testfile bs=4k count=5000", "r");
+	(void) f;
+
+	signal(SIGINT, int_exit);
+
+	if (fork() == 0) {
+		read_trace_pipe();
+	} else {
+		sleep(120);
+		print_hist(map_fd[0]);
+	}
+	return 0;
+}
-- 
2.2.0.33.gc18b867

--
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/