linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Adrian Hunter <adrian.hunter@intel.com>
To: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>, Andi Kleen <ak@linux.intel.com>,
	linux-kernel@vger.kernel.org,
	Adrian Hunter <adrian.hunter@intel.com>
Subject: [PATCH 09/16] perf thread-stack: Add thread_stack__sample_late()
Date: Wed,  1 Apr 2020 13:16:06 +0300	[thread overview]
Message-ID: <20200401101613.6201-10-adrian.hunter@intel.com> (raw)
In-Reply-To: <20200401101613.6201-1-adrian.hunter@intel.com>

Add a thread stack function to create a call chain for hardware events
where the sample records get created some time after the event occurred.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
---
 tools/perf/util/thread-stack.c | 57 ++++++++++++++++++++++++++++++++++
 tools/perf/util/thread-stack.h |  3 ++
 2 files changed, 60 insertions(+)

diff --git a/tools/perf/util/thread-stack.c b/tools/perf/util/thread-stack.c
index 0885967d5bc3..83f6c83f5617 100644
--- a/tools/perf/util/thread-stack.c
+++ b/tools/perf/util/thread-stack.c
@@ -497,6 +497,63 @@ void thread_stack__sample(struct thread *thread, int cpu,
 	chain->nr = i;
 }
 
+/*
+ * Hardware sample records, created some time after the event occurred, need to
+ * have subsequent addresses removed from the call chain.
+ */
+void thread_stack__sample_late(struct thread *thread, int cpu,
+			       struct ip_callchain *chain, size_t sz,
+			       u64 sample_ip, u64 kernel_start)
+{
+	struct thread_stack *ts = thread__stack(thread, cpu);
+	u64 sample_context = callchain_context(sample_ip, kernel_start);
+	u64 last_context, context, ip;
+	size_t nr = 0, j;
+
+	if (sz < 2) {
+		chain->nr = 0;
+		return;
+	}
+
+	if (!ts)
+		goto out;
+
+	/*
+	 * When tracing kernel space, kernel addresses occur at the top of the
+	 * call chain after the event occurred but before tracing stopped.
+	 * Skip them.
+	 */
+	for (j = 1; j <= ts->cnt; j++) {
+		ip = ts->stack[ts->cnt - j].ret_addr;
+		context = callchain_context(ip, kernel_start);
+		if (context == PERF_CONTEXT_USER ||
+		    (context == sample_context && ip == sample_ip))
+			break;
+	}
+
+	last_context = sample_ip; /* Use sample_ip as an invalid context */
+
+	for (; nr < sz && j <= ts->cnt; nr++, j++) {
+		ip = ts->stack[ts->cnt - j].ret_addr;
+		context = callchain_context(ip, kernel_start);
+		if (context != last_context) {
+			if (nr >= sz - 1)
+				break;
+			chain->ips[nr++] = context;
+			last_context = context;
+		}
+		chain->ips[nr] = ip;
+	}
+out:
+	if (nr) {
+		chain->nr = nr;
+	} else {
+		chain->ips[0] = sample_context;
+		chain->ips[1] = sample_ip;
+		chain->nr = 2;
+	}
+}
+
 struct call_return_processor *
 call_return_processor__new(int (*process)(struct call_return *cr, u64 *parent_db_id, void *data),
 			   void *data)
diff --git a/tools/perf/util/thread-stack.h b/tools/perf/util/thread-stack.h
index e1ec5a58f1b2..8962ddc4e1ab 100644
--- a/tools/perf/util/thread-stack.h
+++ b/tools/perf/util/thread-stack.h
@@ -85,6 +85,9 @@ int thread_stack__event(struct thread *thread, int cpu, u32 flags, u64 from_ip,
 void thread_stack__set_trace_nr(struct thread *thread, int cpu, u64 trace_nr);
 void thread_stack__sample(struct thread *thread, int cpu, struct ip_callchain *chain,
 			  size_t sz, u64 ip, u64 kernel_start);
+void thread_stack__sample_late(struct thread *thread, int cpu,
+			       struct ip_callchain *chain, size_t sz, u64 ip,
+			       u64 kernel_start);
 int thread_stack__flush(struct thread *thread);
 void thread_stack__free(struct thread *thread);
 size_t thread_stack__depth(struct thread *thread, int cpu);
-- 
2.17.1


  parent reply	other threads:[~2020-04-01 10:18 UTC|newest]

Thread overview: 47+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-04-01 10:15 [PATCH 00/16] perf intel-pt: Sampling improvements Adrian Hunter
2020-04-01 10:15 ` [PATCH 01/16] perf auxtrace: Add ->evsel_is_auxtrace() callback Adrian Hunter
2020-04-22 12:17   ` [tip: perf/core] " tip-bot2 for Adrian Hunter
2020-04-01 10:15 ` [PATCH 02/16] perf intel-pt: Implement " Adrian Hunter
2020-04-22 12:17   ` [tip: perf/core] " tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 03/16] perf intel-bts: " Adrian Hunter
2020-04-22 12:17   ` [tip: perf/core] " tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 04/16] perf arm-spe: " Adrian Hunter
2020-04-02  3:03   ` Leo Yan
2020-04-22 12:17   ` [tip: perf/core] " tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 05/16] perf cs-etm: " Adrian Hunter
2020-04-01 17:11   ` Mathieu Poirier
2020-04-22 12:17   ` [tip: perf/core] " tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 06/16] perf s390-cpumsf: " Adrian Hunter
2020-04-01 14:10   ` Thomas Richter
2020-04-22 12:17   ` [tip: perf/core] " tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 07/16] perf auxtrace: For reporting purposes, un-group AUX area event Adrian Hunter
2020-04-22 12:17   ` [tip: perf/core] " tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 08/16] perf auxtrace: Add an option to synthesize callchains for regular events Adrian Hunter
2020-04-22 12:17   ` [tip: perf/core] " tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` Adrian Hunter [this message]
2020-04-22 12:17   ` [tip: perf/core] perf thread-stack: Add thread_stack__sample_late() tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 10/16] perf tools: Add support for synthesized sample type Adrian Hunter
2020-04-16 14:54   ` Arnaldo Carvalho de Melo
2020-04-16 14:57     ` Arnaldo Carvalho de Melo
2020-04-16 15:01       ` Arnaldo Carvalho de Melo
2020-04-22 12:17   ` [tip: perf/core] perf evsel: Be consistent when looking which evsel PERF_SAMPLE_ bits are set tip-bot2 for Adrian Hunter
2020-04-22 12:17   ` [tip: perf/core] perf evsel: Add support for synthesized sample type tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 11/16] perf intel-pt: Add support for synthesizing callchains for regular events Adrian Hunter
2020-04-16 15:14   ` Arnaldo Carvalho de Melo
2020-04-17 13:50     ` Adrian Hunter
2020-04-17 21:37       ` Arnaldo Carvalho de Melo
2020-04-20  3:04         ` Andi Kleen
2020-04-22 12:17   ` [tip: perf/core] " tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 12/16] perf tools: Move and globalize perf_evsel__find_pmu() and perf_evsel__is_aux_event() Adrian Hunter
2020-04-18 11:50   ` Arnaldo Carvalho de Melo
2020-04-18 12:04     ` Arnaldo Carvalho de Melo
2020-04-22 12:17   ` [tip: perf/core] perf evsel: " tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 13/16] perf tools: Move leader-sampling configuration Adrian Hunter
2020-04-16 15:29   ` Arnaldo Carvalho de Melo
2020-04-22 12:17   ` [tip: perf/core] perf evlist: " tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 14/16] perf tools: Rearrange perf_evsel__config_leader_sampling() Adrian Hunter
2020-04-22 12:17   ` [tip: perf/core] perf evsel: " tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 15/16] perf tools: Allow multiple read formats Adrian Hunter
2020-04-22 12:17   ` [tip: perf/core] perf evlist: " tip-bot2 for Adrian Hunter
2020-04-01 10:16 ` [PATCH 16/16] perf tools: Add support for leader-sampling with AUX area events Adrian Hunter
2020-04-22 12:17   ` [tip: perf/core] " tip-bot2 for Adrian Hunter

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=20200401101613.6201-10-adrian.hunter@intel.com \
    --to=adrian.hunter@intel.com \
    --cc=acme@kernel.org \
    --cc=ak@linux.intel.com \
    --cc=jolsa@redhat.com \
    --cc=linux-kernel@vger.kernel.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).