Merge branch 'Add bpf_loop helper'

Joanne Koong says:

====================

This patchset add a new helper, bpf_loop.

One of the complexities of using for loops in bpf programs is that the verifier
needs to ensure that in every possibility of the loop logic, the loop will always
terminate. As such, there is a limit on how many iterations the loop can do.

The bpf_loop helper moves the loop logic into the kernel and can thereby
guarantee that the loop will always terminate. The bpf_loop helper simplifies
a lot of the complexity the verifier needs to check, as well as removes the
constraint on the number of loops able to be run.

From the test results, we see that using bpf_loop in place
of the traditional for loop led to a decrease in verification time
and number of bpf instructions by ~99%. The benchmark results show
that as the number of iterations increases, the overhead per iteration
decreases.

The high-level overview of the patches -
Patch 1 - kernel-side + API changes for adding bpf_loop
Patch 2 - tests
Patch 3 - use bpf_loop in strobemeta + pyperf600 and measure verifier performance
Patch 4 - benchmark for throughput + latency of bpf_loop call

v3 -> v4:
~ Address nits: use usleep for triggering bpf programs, fix copyright style

v2 -> v3:
~ Rerun benchmarks on physical machine, update results
~ Propagate original error codes in the verifier

v1 -> v2:
~ Change helper name to bpf_loop (instead of bpf_for_each)
~ Set max nr_loops (~8 million loops) for bpf_loop call
~ Split tests + strobemeta/pyperf600 changes into two patches
~ Add new ops_report_final helper for outputting throughput and latency
====================

Signed-off-by: Alexei Starovoitov <ast@kernel.org>

1 files changed, 112 insertions, 0 deletions

diff --git a/tools/testing/selftests/bpf/progs/bpf_loop.c b/tools/testing/selftests/bpf/progs/bpf_loop.c
new file mode 100644
index 000000000000..12349e4601e8
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/bpf_loop.c
@@ -0,0 +1,112 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2021 Facebook */
+
+#include "vmlinux.h"
+#include <bpf/bpf_helpers.h>
+
+char _license[] SEC("license") = "GPL";
+
+struct callback_ctx {
+	int output;
+};
+
+/* These should be set by the user program */
+u32 nested_callback_nr_loops;
+u32 stop_index = -1;
+u32 nr_loops;
+int pid;
+
+/* Making these global variables so that the userspace program
+ * can verify the output through the skeleton
+ */
+int nr_loops_returned;
+int g_output;
+int err;
+
+static int callback(__u32 index, void *data)
+{
+	struct callback_ctx *ctx = data;
+
+	if (index >= stop_index)
+		return 1;
+
+	ctx->output += index;
+
+	return 0;
+}
+
+static int empty_callback(__u32 index, void *data)
+{
+	return 0;
+}
+
+static int nested_callback2(__u32 index, void *data)
+{
+	nr_loops_returned += bpf_loop(nested_callback_nr_loops, callback, data, 0);
+
+	return 0;
+}
+
+static int nested_callback1(__u32 index, void *data)
+{
+	bpf_loop(nested_callback_nr_loops, nested_callback2, data, 0);
+	return 0;
+}
+
+SEC("fentry/__x64_sys_nanosleep")
+int test_prog(void *ctx)
+{
+	struct callback_ctx data = {};
+
+	if (bpf_get_current_pid_tgid() >> 32 != pid)
+		return 0;
+
+	nr_loops_returned = bpf_loop(nr_loops, callback, &data, 0);
+
+	if (nr_loops_returned < 0)
+		err = nr_loops_returned;
+	else
+		g_output = data.output;
+
+	return 0;
+}
+
+SEC("fentry/__x64_sys_nanosleep")
+int prog_null_ctx(void *ctx)
+{
+	if (bpf_get_current_pid_tgid() >> 32 != pid)
+		return 0;
+
+	nr_loops_returned = bpf_loop(nr_loops, empty_callback, NULL, 0);
+
+	return 0;
+}
+
+SEC("fentry/__x64_sys_nanosleep")
+int prog_invalid_flags(void *ctx)
+{
+	struct callback_ctx data = {};
+
+	if (bpf_get_current_pid_tgid() >> 32 != pid)
+		return 0;
+
+	err = bpf_loop(nr_loops, callback, &data, 1);
+
+	return 0;
+}
+
+SEC("fentry/__x64_sys_nanosleep")
+int prog_nested_calls(void *ctx)
+{
+	struct callback_ctx data = {};
+
+	if (bpf_get_current_pid_tgid() >> 32 != pid)
+		return 0;
+
+	nr_loops_returned = 0;
+	bpf_loop(nr_loops, nested_callback1, &data, 0);
+
+	g_output = data.output;
+
+	return 0;
+}