From: Arnaldo Carvalho de Melo <acme@infradead.org>
To: Ingo Molnar <mingo@elte.hu>
Cc: linux-kernel@vger.kernel.org,
Arnaldo Carvalho de Melo <acme@redhat.com>,
David Ahern <dsahern@gmail.com>,
Frederic Weisbecker <fweisbec@gmail.com>,
Mike Galbraith <efault@gmx.de>, Paul Mackerras <paulus@samba.org>,
Peter Zijlstra <peterz@infradead.org>,
Stephane Eranian <eranian@google.com>
Subject: [PATCH 04/15] perf hists: Threaded addition and sorting of entries
Date: Thu, 6 Oct 2011 14:22:27 -0300 [thread overview]
Message-ID: <1317921758-4154-5-git-send-email-acme@infradead.org> (raw)
In-Reply-To: <1317921758-4154-1-git-send-email-acme@infradead.org>
From: Arnaldo Carvalho de Melo <acme@redhat.com>
By using a mutex just for inserting and rotating two hist_entry rb
trees, so that when sorting we can get the last batch of entries created
from the ring buffer, merge it with whatever we have processed so far
and show the output while new entries are being added.
The 'report' tool continues, for now, to do it without threading, but
will use this in the future to allow visualization of results in long
perf.data sessions while the entries are being processed.
The new 'top' tool will be the first user.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-9b05atsn0q6m7fqgrug8fk2i@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
tools/perf/util/evsel.c | 1 +
tools/perf/util/hist.c | 111 ++++++++++++++++++++++++++++++++++-------------
tools/perf/util/hist.h | 9 ++++
tools/perf/util/sort.h | 1 +
4 files changed, 92 insertions(+), 30 deletions(-)
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index a03a36b..8b58626 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -37,6 +37,7 @@ void perf_evsel__init(struct perf_evsel *evsel,
evsel->idx = idx;
evsel->attr = *attr;
INIT_LIST_HEAD(&evsel->node);
+ hists__init(&evsel->hists);
}
struct perf_evsel *perf_evsel__new(struct perf_event_attr *attr, int idx)
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index 32c9086..80fe30d 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -118,6 +118,7 @@ static void hists__inc_nr_entries(struct hists *hists, struct hist_entry *h)
if (!h->filtered) {
hists__calc_col_len(hists, h);
++hists->nr_entries;
+ hists->stats.total_period += h->period;
}
}
@@ -132,7 +133,7 @@ struct hist_entry *__hists__add_entry(struct hists *hists,
struct addr_location *al,
struct symbol *sym_parent, u64 period)
{
- struct rb_node **p = &hists->entries.rb_node;
+ struct rb_node **p;
struct rb_node *parent = NULL;
struct hist_entry *he;
struct hist_entry entry = {
@@ -150,9 +151,13 @@ struct hist_entry *__hists__add_entry(struct hists *hists,
};
int cmp;
+ pthread_mutex_lock(&hists->lock);
+
+ p = &hists->entries_in->rb_node;
+
while (*p != NULL) {
parent = *p;
- he = rb_entry(parent, struct hist_entry, rb_node);
+ he = rb_entry(parent, struct hist_entry, rb_node_in);
cmp = hist_entry__cmp(&entry, he);
@@ -170,12 +175,14 @@ struct hist_entry *__hists__add_entry(struct hists *hists,
he = hist_entry__new(&entry);
if (!he)
- return NULL;
- rb_link_node(&he->rb_node, parent, p);
- rb_insert_color(&he->rb_node, &hists->entries);
- hists__inc_nr_entries(hists, he);
+ goto out_unlock;
+
+ rb_link_node(&he->rb_node_in, parent, p);
+ rb_insert_color(&he->rb_node_in, hists->entries_in);
out:
hist_entry__add_cpumode_period(he, al->cpumode, period);
+out_unlock:
+ pthread_mutex_unlock(&hists->lock);
return he;
}
@@ -233,7 +240,7 @@ static bool hists__collapse_insert_entry(struct hists *hists,
while (*p != NULL) {
parent = *p;
- iter = rb_entry(parent, struct hist_entry, rb_node);
+ iter = rb_entry(parent, struct hist_entry, rb_node_in);
cmp = hist_entry__collapse(iter, he);
@@ -254,35 +261,57 @@ static bool hists__collapse_insert_entry(struct hists *hists,
p = &(*p)->rb_right;
}
- rb_link_node(&he->rb_node, parent, p);
- rb_insert_color(&he->rb_node, root);
+ rb_link_node(&he->rb_node_in, parent, p);
+ rb_insert_color(&he->rb_node_in, root);
return true;
}
-void hists__collapse_resort(struct hists *hists)
+static struct rb_root *hists__get_rotate_entries_in(struct hists *hists)
{
- struct rb_root tmp;
+ struct rb_root *root;
+
+ pthread_mutex_lock(&hists->lock);
+
+ root = hists->entries_in;
+ if (++hists->entries_in > &hists->entries_in_array[1])
+ hists->entries_in = &hists->entries_in_array[0];
+
+ pthread_mutex_unlock(&hists->lock);
+
+ return root;
+}
+
+static void __hists__collapse_resort(struct hists *hists, bool threaded)
+{
+ struct rb_root *root;
struct rb_node *next;
struct hist_entry *n;
- if (!sort__need_collapse)
+ if (!sort__need_collapse && !threaded)
return;
- tmp = RB_ROOT;
- next = rb_first(&hists->entries);
- hists->nr_entries = 0;
- hists__reset_col_len(hists);
+ root = hists__get_rotate_entries_in(hists);
+ next = rb_first(root);
+ hists->stats.total_period = 0;
while (next) {
- n = rb_entry(next, struct hist_entry, rb_node);
- next = rb_next(&n->rb_node);
+ n = rb_entry(next, struct hist_entry, rb_node_in);
+ next = rb_next(&n->rb_node_in);
- rb_erase(&n->rb_node, &hists->entries);
- if (hists__collapse_insert_entry(hists, &tmp, n))
+ rb_erase(&n->rb_node_in, root);
+ if (hists__collapse_insert_entry(hists, &hists->entries_collapsed, n))
hists__inc_nr_entries(hists, n);
}
+}
- hists->entries = tmp;
+void hists__collapse_resort(struct hists *hists)
+{
+ return __hists__collapse_resort(hists, false);
+}
+
+void hists__collapse_resort_threaded(struct hists *hists)
+{
+ return __hists__collapse_resort(hists, true);
}
/*
@@ -315,31 +344,43 @@ static void __hists__insert_output_entry(struct rb_root *entries,
rb_insert_color(&he->rb_node, entries);
}
-void hists__output_resort(struct hists *hists)
+static void __hists__output_resort(struct hists *hists, bool threaded)
{
- struct rb_root tmp;
+ struct rb_root *root;
struct rb_node *next;
struct hist_entry *n;
u64 min_callchain_hits;
min_callchain_hits = hists->stats.total_period * (callchain_param.min_percent / 100);
- tmp = RB_ROOT;
- next = rb_first(&hists->entries);
+ if (sort__need_collapse || threaded)
+ root = &hists->entries_collapsed;
+ else
+ root = hists->entries_in;
+
+ next = rb_first(root);
+ hists->entries = RB_ROOT;
hists->nr_entries = 0;
hists__reset_col_len(hists);
while (next) {
- n = rb_entry(next, struct hist_entry, rb_node);
- next = rb_next(&n->rb_node);
+ n = rb_entry(next, struct hist_entry, rb_node_in);
+ next = rb_next(&n->rb_node_in);
- rb_erase(&n->rb_node, &hists->entries);
- __hists__insert_output_entry(&tmp, n, min_callchain_hits);
+ __hists__insert_output_entry(&hists->entries, n, min_callchain_hits);
hists__inc_nr_entries(hists, n);
}
+}
- hists->entries = tmp;
+void hists__output_resort(struct hists *hists)
+{
+ return __hists__output_resort(hists, false);
+}
+
+void hists__output_resort_threaded(struct hists *hists)
+{
+ return __hists__output_resort(hists, true);
}
static size_t callchain__fprintf_left_margin(FILE *fp, int left_margin)
@@ -1043,3 +1084,13 @@ size_t hists__fprintf_nr_events(struct hists *hists, FILE *fp)
return ret;
}
+
+void hists__init(struct hists *hists)
+{
+ memset(hists, 0, sizeof(*hists));
+ hists->entries_in_array[0] = hists->entries_in_array[1] = RB_ROOT;
+ hists->entries_in = &hists->entries_in_array[0];
+ hists->entries_collapsed = RB_ROOT;
+ hists->entries = RB_ROOT;
+ pthread_mutex_init(&hists->lock, NULL);
+}
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index 861ffc3..d3f976c 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -2,6 +2,7 @@
#define __PERF_HIST_H
#include <linux/types.h>
+#include <pthread.h>
#include "callchain.h"
extern struct callchain_param callchain_param;
@@ -43,8 +44,12 @@ enum hist_column {
};
struct hists {
+ struct rb_root entries_in_array[2];
+ struct rb_root *entries_in;
struct rb_root entries;
+ struct rb_root entries_collapsed;
u64 nr_entries;
+ pthread_mutex_t lock;
struct events_stats stats;
u64 event_stream;
u16 col_len[HISTC_NR_COLS];
@@ -52,6 +57,8 @@ struct hists {
struct callchain_cursor callchain_cursor;
};
+void hists__init(struct hists *hists);
+
struct hist_entry *__hists__add_entry(struct hists *self,
struct addr_location *al,
struct symbol *parent, u64 period);
@@ -67,7 +74,9 @@ int hist_entry__snprintf(struct hist_entry *self, char *bf, size_t size,
void hist_entry__free(struct hist_entry *);
void hists__output_resort(struct hists *self);
+void hists__output_resort_threaded(struct hists *hists);
void hists__collapse_resort(struct hists *self);
+void hists__collapse_resort_threaded(struct hists *hists);
void hists__inc_nr_events(struct hists *self, u32 type);
size_t hists__fprintf_nr_events(struct hists *self, FILE *fp);
diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h
index 77d0388..03851e3 100644
--- a/tools/perf/util/sort.h
+++ b/tools/perf/util/sort.h
@@ -45,6 +45,7 @@ extern enum sort_type sort__first_dimension;
* @nr_rows - rows expanded in callchain, recalculated on folding/unfolding
*/
struct hist_entry {
+ struct rb_node rb_node_in;
struct rb_node rb_node;
u64 period;
u64 period_sys;
--
1.6.2.5
next prev parent reply other threads:[~2011-10-06 17:27 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-10-06 17:22 [GIT PULL 00/15] perf/core improvements Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 01/15] perf hists: Stop using 'self' for struct hists Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 02/15] perf hists: Allow limiting the number of rows and columns in fprintf Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 03/15] perf report: Add option to show total period Arnaldo Carvalho de Melo
2011-10-06 17:22 ` Arnaldo Carvalho de Melo [this message]
2011-10-06 17:22 ` [PATCH 05/15] perf browsers: Add live mode to the hists, annotate browsers Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 06/15] perf top: Reuse the 'report' hist_entry/hists classes Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 07/15] perf top: Add callgraph support Arnaldo Carvalho de Melo
2011-10-06 21:47 ` David Ahern
2011-10-06 23:02 ` Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 08/15] perf annotate browser: Allow navigation to called functions Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 09/15] perf top: Use the TUI interface by default Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 10/15] perf tools: Fix broken number of samples for perf report -n Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 11/15] perf annotate browser: Use -> to navigate on assembly lines Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 12/15] perf hists browser: Don't offer symbol actions when symbols not on --sort Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 13/15] perf hists browser: Fix TAB/UNTAB use with multiple events Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 14/15] perf hists browser: Update the browser.nr_entries after the timer Arnaldo Carvalho de Melo
2011-10-06 17:22 ` [PATCH 15/15] perf tools: Make perf.data more self-descriptive (v8) Arnaldo Carvalho de Melo
2011-10-10 5:11 ` [GIT PULL 00/15] perf/core improvements Ingo Molnar
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=1317921758-4154-5-git-send-email-acme@infradead.org \
--to=acme@infradead.org \
--cc=acme@redhat.com \
--cc=dsahern@gmail.com \
--cc=efault@gmx.de \
--cc=eranian@google.com \
--cc=fweisbec@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=paulus@samba.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
all inboxes | Powered by JetHome®