offwaketime: Add support for --state (#2940)

Since offwaketime is really an amalgamation of offcputime and wakeuptime
there is no reason why it shouldn't support the --state argument of the
former.

Co-authored-by: Nikolay Borisov <nborisov@suse.com>

man/man8/offwaketime.8

@@ -2,7 +2,7 @@
 .SH NAME
 offwaketime \- Summarize blocked time by off-CPU stack + waker stack. Uses Linux eBPF/bcc.
 .SH SYNOPSIS
-.B offwaketime [\-h] [\-p PID | \-t TID | \-u | \-k] [\-U | \-K] [\-f] [\-\-stack-storage-size STACK_STORAGE_SIZE] [\-m MIN_BLOCK_TIME] [\-M MAX_BLOCK_TIME] [duration]
+.B offwaketime [\-h] [\-p PID | \-t TID | \-u | \-k] [\-U | \-K] [\-f] [\-\-stack-storage-size STACK_STORAGE_SIZE] [\-m MIN_BLOCK_TIME] [\-M MAX_BLOCK_TIME] [\-\-state STATE] [duration]
 .SH DESCRIPTION
 This program shows kernel stack traces and task names that were blocked and
 "off-CPU", along with the stack traces and task names for the threads that woke
@@ -64,6 +64,10 @@ The amount of time in microseconds over which we store traces (default 1)
 .TP
 \-M MAX_BLOCK_TIME
 The amount of time in microseconds under which we store traces (default U64_MAX)
+.TP
+\-\-state
+Filter on this thread state bitmask (eg, 2 == TASK_UNINTERRUPTIBLE).
+See include/linux/sched.h for states.
 .SH EXAMPLES
 .TP
 Trace all thread blocking events, and summarize (in-kernel) by user and kernel off-CPU stack trace, waker stack traces, task names, and total blocked time:

tools/offwaketime.py

@@ -93,6 +93,9 @@ def stack_id_err(stack_id):
     type=positive_nonzero_int,
     help="the amount of time in microseconds under which we " +
          "store traces (default U64_MAX)")
+parser.add_argument("--state", type=positive_int,
+    help="filter on this thread state bitmask (eg, 2 == TASK_UNINTERRUPTIBLE" +
+         ") see include/linux/sched.h")
 parser.add_argument("--ebpf", action="store_true",
     help=argparse.SUPPRESS)
 args = parser.parse_args()
@@ -147,7 +150,7 @@ def signal_ignore(signal, frame):
     u32 pid = p->pid;
     u32 tgid = p->tgid;
 
-    if (!(THREAD_FILTER)) {
+    if (!((THREAD_FILTER) && (STATE_FILTER))) {
         return 0;
     }
 
@@ -171,7 +174,7 @@ def signal_ignore(signal, frame):
     u64 ts = bpf_ktime_get_ns();
 
     // Record timestamp for the previous Process (Process going into waiting)
-    if (THREAD_FILTER) {
+    if ((THREAD_FILTER) && (STATE_FILTER)) {
         start.update(&pid, &ts);
     }
 
@@ -234,7 +237,15 @@ def signal_ignore(signal, frame):
 else:
     thread_context = "all threads"
     thread_filter = '1'
+if args.state == 0:
+    state_filter = 'p->state == 0'
+elif args.state:
+    # these states are sometimes bitmask checked
+    state_filter = 'p->state & %d' % args.state
+else:
+    state_filter = '1'
 bpf_text = bpf_text.replace('THREAD_FILTER', thread_filter)
+bpf_text = bpf_text.replace('STATE_FILTER', state_filter)
 
 # set stack storage size
 bpf_text = bpf_text.replace('STACK_STORAGE_SIZE', str(args.stack_storage_size))

tools/offwaketime_example.txt

@@ -307,7 +307,7 @@ USAGE message:
 # ./offwaketime -h
 usage: offwaketime [-h] [-p PID | -t TID | -u | -k] [-U | -K] [-d] [-f]
                    [--stack-storage-size STACK_STORAGE_SIZE]
-                   [-m MIN_BLOCK_TIME] [-M MAX_BLOCK_TIME]
+                   [-m MIN_BLOCK_TIME] [-M MAX_BLOCK_TIME] [--state STATE]
                    [duration]
 
 Summarize blocked time by kernel stack trace + waker stack
@@ -340,6 +340,8 @@ optional arguments:
   -M MAX_BLOCK_TIME, --max-block-time MAX_BLOCK_TIME
                         the amount of time in microseconds under which we
                         store traces (default U64_MAX)
+  --state STATE         filter on this thread state bitmask (eg, 2 ==
+                        TASK_UNINTERRUPTIBLE) see include/linux/sched.h
 
 examples:
     ./offwaketime             # trace off-CPU + waker stack time until Ctrl-C