tools: add option --cgroupmap to profile.py

man/man8/profile.8

@@ -1,9 +1,9 @@
-.TH profile 8 "2016-07-17" "USER COMMANDS"
+.TH profile 8 "2020-03-18" "USER COMMANDS"
 .SH NAME
 profile \- Profile CPU usage by sampling stack traces. Uses Linux eBPF/bcc.
 .SH SYNOPSIS
 .B profile [\-adfh] [\-p PID | \-L TID] [\-U | \-K] [\-F FREQUENCY | \-c COUNT]
-.B [\-\-stack\-storage\-size COUNT] [duration]
+.B [\-\-stack\-storage\-size COUNT] [\-\-cgroupmap CGROUPMAP] [duration]
 .SH DESCRIPTION
 This is a CPU profiler. It works by taking samples of stack traces at timed
 intervals. It will help you understand and quantify CPU usage: which code is
@@ -62,6 +62,9 @@ The maximum number of unique stack traces that the kernel will count (default
 \-C cpu
 Collect stacks only from specified cpu.
 .TP
+\-\-cgroupmap MAPPATH
+Profile cgroups in this BPF map only (filtered in-kernel).
+.TP
 duration
 Duration to trace, in seconds.
 .SH EXAMPLES
@@ -97,6 +100,10 @@ Profile for 5 seconds and output in folded stack format (suitable as input for f
 Profile kernel stacks only:
 #
 .B profile -K
+.TP
+Profile a set of cgroups only (see filtering_by_cgroups.md from bcc sources for more details):
+#
+.B profile \-\-cgroupmap /sys/fs/bpf/test01
 .SH DEBUGGING
 See "[unknown]" frames with bogus addresses? This can happen for different
 reasons. Your best approach is to get Linux perf to work first, and then to

tools/profile.py

@@ -72,6 +72,7 @@ def stack_id_err(stack_id):
  ./profile -L 185 # only profile thread with TID 185
  ./profile -U # only show user space stacks (no kernel)
  ./profile -K # only show kernel space stacks (no user)
+ ./profile --cgroupmap ./mappath # only trace cgroups in this BPF map
 """
 parser = argparse.ArgumentParser(
  description="Profile CPU stack traces at a timed interval",
@@ -112,6 +113,8 @@ def stack_id_err(stack_id):
  help="cpu number to run profile on")
 parser.add_argument("--ebpf", action="store_true",
  help=argparse.SUPPRESS)
+parser.add_argument("--cgroupmap",
+ help="trace cgroups in this BPF map only")
 
 # option logic
 args = parser.parse_args()
@@ -143,6 +146,10 @@ def stack_id_err(stack_id):
 BPF_HASH(counts, struct key_t);
 BPF_STACK_TRACE(stack_traces, STACK_STORAGE_SIZE);
 
+#if CGROUPSET
+BPF_TABLE_PINNED("hash", u64, u64, cgroupset, 1024, "CGROUPPATH");
+#endif
+
 // This code gets a bit complex. Probably not suitable for casual hacking.
 
 int do_perf_event(struct bpf_perf_event_data *ctx) {
@@ -156,6 +163,13 @@ def stack_id_err(stack_id):
  if (!(THREAD_FILTER))
  return 0;
 
+#if CGROUPSET
+ u64 cgroupid = bpf_get_current_cgroup_id();
+ if (cgroupset.lookup(&cgroupid) == NULL) {
+ return 0;
+ }
+#endif
+
  // create map key
  struct key_t key = {.pid = tgid};
  bpf_get_current_comm(&key.name, sizeof(key.name));
@@ -232,6 +246,11 @@ def stack_id_err(stack_id):
  stack_context = "user + kernel"
 bpf_text = bpf_text.replace('USER_STACK_GET', user_stack_get)
 bpf_text = bpf_text.replace('KERNEL_STACK_GET', kernel_stack_get)
+if args.cgroupmap:
+ bpf_text = bpf_text.replace('CGROUPSET', '1')
+ bpf_text = bpf_text.replace('CGROUPPATH', args.cgroupmap)
+else:
+ bpf_text = bpf_text.replace('CGROUPSET', '0')
 
 sample_freq = 0
 sample_period = 0

tools/profile_example.txt

@@ -703,12 +703,21 @@ WARNING: 8 stack traces could not be displayed. Consider increasing --stack-stor
 
 Run ./profile -h to see the default.
 
+The --cgroupmap option filters based on a cgroup set. It is meant to be used
+with an externally created map.
+
+# ./profile --cgroupmap /sys/fs/bpf/test01
+
+For more details, see docs/filtering_by_cgroups.md
+
 
 USAGE message:
 
 # ./profile -h
-usage: profile.py [-h] [-p PID | -L TID] [-U | -K] [-F FREQUENCY | -c COUNT] [-d] [-a]
- [-I] [-f] [--stack-storage-size STACK_STORAGE_SIZE] [-C CPU]
+usage: profile.py [-h] [-p PID | -L TID] [-U | -K] [-F FREQUENCY | -c COUNT]
+ [-d] [-a] [-I] [-f]
+ [--stack-storage-size STACK_STORAGE_SIZE] [-C CPU]
+ [--cgroupmap CGROUPMAP]
  [duration]
 
 Profile CPU stack traces at a timed interval
@@ -739,6 +748,8 @@ optional arguments:
  the number of unique stack traces that can be stored
  and displayed (default 16384)
  -C CPU, --cpu CPU cpu number to run profile on
+ --cgroupmap CGROUPMAP
+ trace cgroups in this BPF map only
 
 examples:
  ./profile # profile stack traces at 49 Hertz until Ctrl-C
@@ -750,3 +761,4 @@ examples:
  ./profile -L 185 # only profile thread with TID 185
  ./profile -U # only show user space stacks (no kernel)
  ./profile -K # only show kernel space stacks (no user)
+ ./profile --cgroupmap ./mappath # only trace cgroups in this BPF map