hardirqs: add --count for event counts (#1460)

man/man8/hardirqs.8

@@ -2,7 +2,7 @@
 .SH NAME
 hardirqs \- Measure hard IRQ (hard interrupt) event time. Uses Linux eBPF/bcc.
 .SH SYNOPSIS
-.B hardirqs [\-h] [\-T] [\-N] [\-d] [interval] [count]
+.B hardirqs [\-h] [\-T] [\-N] [\-C] [\-d] [interval] [outputs]
 .SH DESCRIPTION
 This summarizes the time spent servicing hard IRQs (hard interrupts), and can
 show this time as either totals or histogram distributions. A system-wide
@@ -25,10 +25,13 @@ Print usage message.
 Include timestamps on output.
 .TP
 \-N
-Output in nanoseconds
+Output in nanoseconds.
+.TP
+\-C
+Count events only.
 .TP
 \-d
-Show IRQ time distribution as histograms
+Show IRQ time distribution as histograms.
 .SH EXAMPLES
 .TP
 Sum hard IRQ event time until Ctrl-C:

tools/hardirqs.py

@@ -4,7 +4,7 @@
 # hardirqs  Summarize hard IRQ (interrupt) event time.
 #           For Linux, uses BCC, eBPF.
 #
-# USAGE: hardirqs [-h] [-T] [-Q] [-m] [-D] [interval] [count]
+# USAGE: hardirqs [-h] [-T] [-N] [-C] [-d] [interval] [outputs]
 #
 # Thanks Amer Ather for help understanding irq behavior.
 #
@@ -33,15 +33,23 @@ parser.add_argument("-T", "--timestamp", action="store_true",
     help="include timestamp on output")
 parser.add_argument("-N", "--nanoseconds", action="store_true",
     help="output in nanoseconds")
+parser.add_argument("-C", "--count", action="store_true",
+    help="show event counts instead of timing")
 parser.add_argument("-d", "--dist", action="store_true",
     help="show distributions as histograms")
 parser.add_argument("interval", nargs="?", default=99999999,
     help="output interval, in seconds")
-parser.add_argument("count", nargs="?", default=99999999,
+parser.add_argument("outputs", nargs="?", default=99999999,
     help="number of outputs")
 args = parser.parse_args()
-countdown = int(args.count)
-if args.nanoseconds:
+countdown = int(args.outputs)
+if args.count and (args.dist or args.nanoseconds):
+    print("The --count option can't be used with time-based options")
+    exit()
+if args.count:
+    factor = 1
+    label = "count"
+elif args.nanoseconds:
     factor = 1
     label = "nsecs"
 else:
@@ -64,6 +72,22 @@ BPF_HASH(start, u32);
 BPF_HASH(irqdesc, u32, struct irq_desc *);
 BPF_HISTOGRAM(dist, irq_key_t);
 
+// count IRQ
+int count_only(struct pt_regs *ctx, struct irq_desc *desc)
+{
+    u32 pid = bpf_get_current_pid_tgid();
+
+    struct irqaction *action = desc->action;
+    char *name = (char *)action->name;
+
+    irq_key_t key = {.slot = 0 /* ignore */};
+    bpf_probe_read(&key.name, sizeof(key.name), name);
+    u64 zero = 0, *vp = dist.lookup_or_init(&key, &zero);
+    (*vp)++;
+
+    return 0;
+}
+
 // time IRQ
 int trace_start(struct pt_regs *ctx, struct irq_desc *desc)
 {
@@ -119,10 +143,14 @@ if debug:
 b = BPF(text=bpf_text)
 
 # these should really use irq:irq_handler_entry/exit tracepoints:
-b.attach_kprobe(event="handle_irq_event_percpu", fn_name="trace_start")
-b.attach_kretprobe(event="handle_irq_event_percpu", fn_name="trace_completion")
-
-print("Tracing hard irq event time... Hit Ctrl-C to end.")
+if args.count:
+    b.attach_kprobe(event="handle_irq_event_percpu", fn_name="count_only")
+    print("Tracing hard irq events... Hit Ctrl-C to end.")
+else:
+    b.attach_kprobe(event="handle_irq_event_percpu", fn_name="trace_start")
+    b.attach_kretprobe(event="handle_irq_event_percpu",
+        fn_name="trace_completion")
+    print("Tracing hard irq event time... Hit Ctrl-C to end.")
 
 # output
 exiting = 0 if args.interval else 1

tools/hardirqs_example.txt

@@ -614,21 +614,55 @@ hardirq = 'resched3'
       2048 -> 4095       : 39       |*                                       |
 
 
+Sometimes you just want counts of events, and don't need the distribution
+of times. You can use the -C or --count option:
+
+# ./hardirqs.py -C
+Tracing hard irq events... Hit Ctrl-C to end.
+^C
+HARDIRQ                    TOTAL_count
+blkif                                2
+callfuncsingle3                      8
+callfuncsingle2                     10
+callfuncsingle1                     18
+resched7                            25
+callfuncsingle6                     25
+callfuncsingle5                     27
+callfuncsingle0                     27
+eth0                                34
+resched2                            40
+resched1                            66
+timer7                              70
+resched6                            71
+resched0                            73
+resched5                            79
+resched4                            90
+timer6                              95
+timer4                             100
+timer1                             109
+timer2                             115
+timer0                             117
+timer3                             123
+resched3                           140
+timer5                             288
+
+
 USAGE message:
 
 # ./hardirqs -h
-usage: hardirqs [-h] [-T] [-N] [-d] [interval] [count]
+usage: hardirqs [-h] [-T] [-N] [-C] [-d] [interval] [outputs]
 
 Summarize hard irq event time as histograms
 
 positional arguments:
   interval           output interval, in seconds
-  count              number of outputs
+  outputs            number of outputs
 
 optional arguments:
   -h, --help         show this help message and exit
   -T, --timestamp    include timestamp on output
   -N, --nanoseconds  output in nanoseconds
+  -C, --count        show event counts instead of timing
   -d, --dist         show distributions as histograms
 
 examples: