Skip to contentSkip to conversation

selftests/bpf: add uprobe triggering overhead benchmarks

PATCH Superseded Andrii Nakryiko 1 participant 1 message Nov 15, 2021

Subsystems (3)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Andrii Nakryiko

author maintainer
#0 lore raw

PATCH 1/1 selftests/bpf: add uprobe triggering overhead benchmarks

Add benchmark to measure overhead of uprobes and uretprobes. Also have a baseline (no uprobe attached) benchmark.

On my dev machine, baseline benchmark can trigger 130M user_target() invocations. When uprobe is attached, this falls to just 700K. With uretprobe, we get down to 520K:

  $ sudo ./bench trig-uprobe-base -a
  Summary: hits  131.289 ± 2.872M/s
  $ sudo ./bench trig-uprobe -a
  Summary: hits    0.700 ± 0.008M/s
  $ sudo ./bench trig-uretprobe -a
  Summary: hits    0.518 ± 0.006M/s

This means that uprobe overhead is around 1.4 microseconds for uprobe and 2 microseconds for uretprobe.

For comparison, just doing a very low-overhead syscall (with no BPF programs attached anywhere) gives:

  $ sudo ./bench trig-base -a
  Summary: hits    4.830 ± 0.036M/s

So uprobes are about 7x slower than pure context switch.

---
 tools/testing/selftests/bpf/Makefile          |  4 +-
 tools/testing/selftests/bpf/bench.c           |  6 ++
 .../selftests/bpf/benchs/bench_trigger.c      | 94 +++++++++++++++++++
 .../selftests/bpf/progs/trigger_bench.c       |  7 ++
 4 files changed, 110 insertions(+), 1 deletion(-)
tools/testing/selftests/bpf/Makefile +3 -1
index 0470802c907c..35684d61aaeb 100644
@@ -533,7 +533,9 @@ $(OUTPUT)/bench_ringbufs.o: $(OUTPUT)/ringbuf_bench.skel.h \
 $(OUTPUT)/bench_bloom_filter_map.o: $(OUTPUT)/bloom_filter_bench.skel.h
 $(OUTPUT)/bench.o: bench.h testing_helpers.h $(BPFOBJ)
 $(OUTPUT)/bench: LDLIBS += -lm
-$(OUTPUT)/bench: $(OUTPUT)/bench.o $(OUTPUT)/testing_helpers.o \
+$(OUTPUT)/bench: $(OUTPUT)/bench.o \
+		 $(OUTPUT)/testing_helpers.o \
+		 $(OUTPUT)/trace_helpers.o \
 		 $(OUTPUT)/bench_count.o \
 		 $(OUTPUT)/bench_rename.o \
 		 $(OUTPUT)/bench_trigger.o \
tools/testing/selftests/bpf/bench.c +6 -0
index cc4722f693e9..10f371376585 100644
@@ -359,6 +359,9 @@ extern const struct bench bench_trig_kprobe;
 extern const struct bench bench_trig_fentry;
 extern const struct bench bench_trig_fentry_sleep;
 extern const struct bench bench_trig_fmodret;
+extern const struct bench bench_trig_uprobe_base;
+extern const struct bench bench_trig_uprobe;
+extern const struct bench bench_trig_uretprobe;
 extern const struct bench bench_rb_libbpf;
 extern const struct bench bench_rb_custom;
 extern const struct bench bench_pb_libbpf;
@@ -385,6 +388,9 @@ static const struct bench *benchs[] = {
 	&bench_trig_fentry,
 	&bench_trig_fentry_sleep,
 	&bench_trig_fmodret,
+	&bench_trig_uprobe_base,
+	&bench_trig_uprobe,
+	&bench_trig_uretprobe,
 	&bench_rb_libbpf,
 	&bench_rb_custom,
 	&bench_pb_libbpf,
tools/testing/selftests/bpf/benchs/bench_trigger.c +94 -0

109 lines not on this page · raw message

tools/testing/selftests/bpf/progs/trigger_bench.c +7 -0
index 9a4d09590b3d..2098f3f27f18 100644
@@ -52,3 +52,10 @@ int bench_trigger_fmodret(void *ctx)
 	__sync_add_and_fetch(&hits, 1);
 	return -22;
 }
+
+SEC("uprobe/self/uprobe_target")
+int bench_trigger_uprobe(void *ctx)
+{
+	__sync_add_and_fetch(&hits, 1);
+	return 0;
+}
Signature · 1 line
2.30.2