From: David Ahern <dsa@cumulusnetworks.com>
To: acme@kernel.org
Cc: mingo@kernel.org, peterz@infradead.org, namhyung@kernel.org,
jolsa@kernel.org, linux-kernel@vger.kernel.org,
David Ahern <dsahern@gmail.com>
Subject: [PATCH v2 5/6] perf kmem: Add option to specify time window of interest
Date: Tue, 29 Nov 2016 10:15:45 -0700 [thread overview]
Message-ID: <1480439746-42695-6-git-send-email-dsahern@gmail.com> (raw)
In-Reply-To: <1480439746-42695-1-git-send-email-dsahern@gmail.com>
Add option to allow user to control analysis window. e.g., collect data
for time window and analyze a segment of interest within that window.
Signed-off-by: David Ahern <dsahern@gmail.com>
---
tools/perf/Documentation/perf-kmem.txt | 7 +++++++
tools/perf/builtin-kmem.c | 24 ++++++++++++++++++++++++
2 files changed, 31 insertions(+)
diff --git a/tools/perf/Documentation/perf-kmem.txt b/tools/perf/Documentation/perf-kmem.txt
index ff0f433b3fce..479fc3261a50 100644
--- a/tools/perf/Documentation/perf-kmem.txt
+++ b/tools/perf/Documentation/perf-kmem.txt
@@ -61,6 +61,13 @@ OPTIONS
default, but this option shows live (currently allocated) pages
instead. (This option works with --page option only)
+--time::
+ Only analyze samples within given time window: <start>,<stop>. Times
+ have the format seconds.microseconds. If start is not given (i.e., time
+ string is ',x.y') then analysis starts at the beginning of the file. If
+ stop time is not given (i.e, time string is 'x.y,') then analysis goes
+ to end of file.
+
SEE ALSO
--------
linkperf:perf-record[1]
diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 7fd6f1e1e293..35a02f8e5a4a 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -11,6 +11,7 @@
#include "util/session.h"
#include "util/tool.h"
#include "util/callchain.h"
+#include "util/time-utils.h"
#include <subcmd/parse-options.h>
#include "util/trace-event.h"
@@ -66,6 +67,10 @@ static struct rb_root root_caller_sorted;
static unsigned long total_requested, total_allocated, total_freed;
static unsigned long nr_allocs, nr_cross_allocs;
+/* filters for controlling start and stop of time of analysis */
+static struct perf_time_interval ptime;
+const char *time_str;
+
static int insert_alloc_stat(unsigned long call_site, unsigned long ptr,
int bytes_req, int bytes_alloc, int cpu)
{
@@ -912,6 +917,15 @@ static int perf_evsel__process_page_free_event(struct perf_evsel *evsel,
return 0;
}
+static bool perf_kmem__skip_sample(struct perf_sample *sample)
+{
+ /* skip sample based on time? */
+ if (perf_time__skip_sample(&ptime, sample->time))
+ return true;
+
+ return false;
+}
+
typedef int (*tracepoint_handler)(struct perf_evsel *evsel,
struct perf_sample *sample);
@@ -931,6 +945,9 @@ static int process_sample_event(struct perf_tool *tool __maybe_unused,
return -1;
}
+ if (perf_kmem__skip_sample(sample))
+ return 0;
+
dump_printf(" ... thread: %s:%d\n", thread__comm_str(thread), thread->tid);
if (evsel->handler != NULL) {
@@ -1894,6 +1911,8 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused)
OPT_CALLBACK_NOOPT(0, "page", NULL, NULL, "Analyze page allocator",
parse_page_opt),
OPT_BOOLEAN(0, "live", &live_page, "Show live page stat"),
+ OPT_STRING(0, "time", &time_str, "str",
+ "Time span of interest (start,stop)"),
OPT_END()
};
const char *const kmem_subcommands[] = { "record", "stat", NULL };
@@ -1954,6 +1973,11 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused)
symbol__init(&session->header.env);
+ if (perf_time__parse_str(&ptime, time_str) != 0) {
+ pr_err("Invalid time string\n");
+ return -EINVAL;
+ }
+
if (!strcmp(argv[0], "stat")) {
setlocale(LC_ALL, "");
--
2.7.4 (Apple Git-66)
next prev parent reply other threads:[~2016-11-29 17:17 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-11-29 17:15 [PATCH v2 0/6] perf: Add option to specify time window of interest David Ahern
2016-11-29 17:15 ` [PATCH v2 1/6] perf tool: Add time-based utility functions David Ahern
2016-12-02 10:41 ` [tip:perf/core] perf tools: " tip-bot for David Ahern
2016-11-29 17:15 ` [PATCH v2 2/6] perf tool: Move parse_nsec_time to time-utils.c David Ahern
2016-12-02 10:41 ` [tip:perf/core] perf tools: " tip-bot for David Ahern
2016-11-29 17:15 ` [PATCH v2 3/6] perf script: Add option to specify time window of interest David Ahern
2016-12-02 10:42 ` [tip:perf/core] " tip-bot for David Ahern
2016-11-29 17:15 ` [PATCH v2 4/6] perf sched timehist: " David Ahern
2016-11-29 18:56 ` Arnaldo Carvalho de Melo
2016-11-29 18:58 ` David Ahern
2016-11-29 19:21 ` Arnaldo Carvalho de Melo
2016-12-02 10:42 ` [tip:perf/core] " tip-bot for David Ahern
2016-11-29 17:15 ` David Ahern [this message]
2016-12-02 10:43 ` [tip:perf/core] perf kmem: " tip-bot for David Ahern
2016-11-29 17:15 ` [PATCH v2 6/6] perf report: " David Ahern
2016-12-02 10:43 ` [tip:perf/core] " tip-bot for David Ahern
2016-11-29 19:15 ` [PATCH v2 0/6] perf: " Arnaldo Carvalho de Melo
2016-11-30 5:26 ` Namhyung Kim
2016-11-30 14:23 ` Arnaldo Carvalho de Melo
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1480439746-42695-6-git-send-email-dsahern@gmail.com \
--to=dsa@cumulusnetworks.com \
--cc=acme@kernel.org \
--cc=dsahern@gmail.com \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=namhyung@kernel.org \
--cc=peterz@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).