perf tools: Create function to perform multiple time range checking
authorJin Yao <yao.jin@linux.intel.com>
Fri, 8 Dec 2017 13:13:44 +0000 (21:13 +0800)
committerArnaldo Carvalho de Melo <acme@redhat.com>
Mon, 8 Jan 2018 14:41:06 +0000 (11:41 -0300)
Previous patch supports the multiple time range.

For example, select the first and second 10% time slices.
perf report --time 10%/1,10%/2

We need a function to check if a timestamp is in the ranges of
[0, 10%) and [10%, 20%].

Note that it includes the last element in [10%, 20%] but it doesn't
include the last element in [0, 10%). It's to avoid the overlap.

This patch implments a new function perf_time__ranges_skip_sample
for this checking.

Change log:

v4: Let perf_time__ranges_skip_sample be compatible with
    perf_time__skip_sample when only one time range.

Signed-off-by: Jin Yao <yao.jin@linux.intel.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1512738826-2628-5-git-send-email-yao.jin@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
tools/perf/util/time-utils.c
tools/perf/util/time-utils.h

index 61c46022de0bf17a88604553f2363c07cf30d472..3f7f18f06982e118c45513d08c56c5c56e4b021d 100644 (file)
@@ -300,6 +300,34 @@ bool perf_time__skip_sample(struct perf_time_interval *ptime, u64 timestamp)
        return false;
 }
 
+bool perf_time__ranges_skip_sample(struct perf_time_interval *ptime_buf,
+                                  int num, u64 timestamp)
+{
+       struct perf_time_interval *ptime;
+       int i;
+
+       if ((timestamp == 0) || (num == 0))
+               return false;
+
+       if (num == 1)
+               return perf_time__skip_sample(&ptime_buf[0], timestamp);
+
+       /*
+        * start/end of multiple time ranges must be valid.
+        */
+       for (i = 0; i < num; i++) {
+               ptime = &ptime_buf[i];
+
+               if (timestamp >= ptime->start &&
+                   ((timestamp < ptime->end && i < num - 1) ||
+                    (timestamp <= ptime->end && i == num - 1))) {
+                       break;
+               }
+       }
+
+       return (i == num) ? true : false;
+}
+
 int timestamp__scnprintf_usec(u64 timestamp, char *buf, size_t sz)
 {
        u64  sec = timestamp / NSEC_PER_SEC;
index 23087231785add817564014378a43460e519d367..34d5eba26bf5b2c9f3dfa6d95d26524826811394 100644 (file)
@@ -18,6 +18,9 @@ int perf_time__percent_parse_str(struct perf_time_interval *ptime_buf, int num,
 
 bool perf_time__skip_sample(struct perf_time_interval *ptime, u64 timestamp);
 
+bool perf_time__ranges_skip_sample(struct perf_time_interval *ptime_buf,
+                                  int num, u64 timestamp);
+
 int timestamp__scnprintf_usec(u64 timestamp, char *buf, size_t sz);
 
 int fetch_current_timestamp(char *buf, size_t sz);