All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Wangnan (F)" <wangnan0@huawei.com>
To: Arnaldo Carvalho de Melo <acme@kernel.org>,
	Namhyung Kim <namhyung@kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: [BUG] perf: memory leak in perf top
Date: Fri, 5 Aug 2016 18:58:12 +0800	[thread overview]
Message-ID: <57A47144.6000406@huawei.com> (raw)

perf top comsumes too much memory than it need.

Using following commands:

  # yes '' > /dev/null
  # perf top -e raw_syscalls:sys_enter

System will quickly become unresponsive because of out of memory.

Using GDB I find two call stacks of malloc. See below.

[call stack 1]

#0 __GI___libc_malloc (bytes=140723951319552) at malloc.c:2841
#1 0x00000000004b7246 in memdup (src=0x7f515e04b054, len=68) at 
../lib/string.c:29
#2 0x00000000004f9299 in hist_entry__init (sample_self=true, 
template=0x7ffcd9213a80, he=0xa37da50) at util/hist.c:401
#3 hist_entry__new (template=template@entry=0x7ffcd9213a80 
<mailto:%28template=template@entry=0x7ffcd9213a80>, 
sample_self=sample_self@entry=true) 
<mailto:sample_self=sample_self@entry=true%29> at util/hist.c:458
#4 0x00000000004fa13f in hists__findnew_entry (hists=0x3549d30, 
hists=0x3549d30, al=0x7ffcd9213ca0, sample_self=true, entry=0x
7ffcd9213a80) at util/hist.c:544
#5 __hists__add_entry (hists=0x3549d30, al=0x7ffcd9213ca0, 
sym_parent=<optimized out>, bi=bi@entry=0x0 <mailto:bi=bi@entry=0x0>, 
mi=mi@entry=0x0 <mailto:mi=mi@entry=0x0>, sampl
e=<optimized out>, sample_self=sample_self@entry=true 
<mailto:sample_self=sample_self@entry=true>, ops=ops@entry=0x0) 
<mailto:ops=ops@entry=0x0%29> at util/hist.c:600
#6 0x00000000004fa3d6 in hists__add_entry (sample_self=true, 
sample=<optimized out>, mi=0x0, bi=0x0, sym_parent=<optimized out
 >, al=<optimized out>, hists=<optimized out>) at util/hist.c:611
#7 iter_add_single_normal_entry (iter=0x7ffcd9213ce0, al=<optimized 
out>) at util/hist.c:823
#8 0x00000000004fc3c5 in hist_entry_iter__add (iter=0x7ffcd9213ce0, 
al=0x7ffcd9213ca0, max_stack_depth=<optimized out>, arg=0x
7ffcd9214120) at util/hist.c:1030
#9 0x00000000004536d7 in perf_event__process_sample (machine=<optimized 
out>, sample=0x7ffcd9213d30, evsel=0x3549bb0, event=<o
ptimized out>, tool=0x7ffcd9214120) at builtin-top.c:800
#10 perf_top__mmap_read_idx (top=top@entry=0x7ffcd9214120 
<mailto:%28top=top@entry=0x7ffcd9214120>, idx=idx@entry=0) 
<mailto:idx=idx@entry=0%29> at builtin-top.c:866
#11 0x00000000004559fd in perf_top__mmap_read (top=0x7ffcd9214120) at 
builtin-top.c:883
#12 __cmd_top (top=0x7ffcd9214120) at builtin-top.c:1018
#13 cmd_top (argc=<optimized out>, argv=<optimized out>, 
prefix=<optimized out>) at builtin-top.c:1344
#14 0x0000000000493561 in run_builtin (p=p@entry=0x948000 
<mailto:%28p=p@entry=0x948000> <commands+288>, argc=argc@entry=4 
<mailto:argc=argc@entry=4>, argv=argv@entry=0x7ffcd92199b0) 
<mailto:argv=argv@entry=0x7ffcd92199b0%29> at
perf.c:357
#15 0x0000000000437157 in handle_internal_command (argv=0x7ffcd92199b0, 
argc=4) at perf.c:419
#16 run_argv (argv=0x7ffcd9219720, argcp=0x7ffcd921972c) at perf.c:465
#17 main (argc=4, argv=0x7ffcd92199b0) at perf.c:609

[call stack 2]

#0 __GI___libc_malloc (bytes=140723951319456) at malloc.c:2841
#1 0x00000000005766de in trace_seq_init (s=0x7ffcd92139b0) at trace-seq.c:60
#2 0x00000000004f7047 in get_trace_output (he=0x7ffcd9213a80) at 
util/sort.c:584
#3 sort__trace_cmp (left=0x40e6d30, right=0x7ffcd9213a80) at util/sort.c:606
#4 0x00000000004fa007 in hist_entry__cmp (right=0x7ffcd9213a80, 
left=0x40e6d30) at util/hist.c:1072
#5 hists__findnew_entry (hists=0x3549d30, hists=0x3549d30, 
al=0x7ffcd9213ca0, sample_self=true, entry=0x7ffcd9213a80) at util/
hist.c:509
#6 __hists__add_entry (hists=0x3549d30, al=0x7ffcd9213ca0, 
sym_parent=<optimized out>, bi=bi@entry=0x0 <mailto:bi=bi@entry=0x0>, 
mi=mi@entry=0x0 <mailto:mi=mi@entry=0x0>, sampl
e=<optimized out>, sample_self=sample_self@entry=true 
<mailto:sample_self=sample_self@entry=true>, ops=ops@entry=0x0) 
<mailto:ops=ops@entry=0x0%29> at util/hist.c:600
#7 0x00000000004fa3d6 in hists__add_entry (sample_self=true, 
sample=<optimized out>, mi=0x0, bi=0x0, sym_parent=<optimized out
 >, al=<optimized out>, hists=<optimized out>) at util/hist.c:611
#8 iter_add_single_normal_entry (iter=0x7ffcd9213ce0, al=<optimized 
out>) at util/hist.c:823
#9 0x00000000004fc3c5 in hist_entry_iter__add (iter=0x7ffcd9213ce0, 
al=0x7ffcd9213ca0, max_stack_depth=<optimized out>, arg=0x
7ffcd9214120) at util/hist.c:1030
#10 0x00000000004536d7 in perf_event__process_sample (machine=<optimized 
out>, sample=0x7ffcd9213d30, evsel=0x3549bb0, event=<o
ptimized out>, tool=0x7ffcd9214120) at builtin-top.c:800
#11 perf_top__mmap_read_idx (top=top@entry=0x7ffcd9214120 
<mailto:%28top=top@entry=0x7ffcd9214120>, idx=idx@entry=0) 
<mailto:idx=idx@entry=0%29> at builtin-top.c:866
#12 0x00000000004559fd in perf_top__mmap_read (top=0x7ffcd9214120) at 
builtin-top.c:883
#13 __cmd_top (top=0x7ffcd9214120) at builtin-top.c:1018
#14 cmd_top (argc=<optimized out>, argv=<optimized out>, 
prefix=<optimized out>) at builtin-top.c:1344
#15 0x0000000000493561 in run_builtin (p=p@entry=0x948000 
<mailto:%28p=p@entry=0x948000> <commands+288>, argc=argc@entry=4 
<mailto:argc=argc@entry=4>, argv=argv@entry=0x7ffcd92199b0) 
<mailto:argv=argv@entry=0x7ffcd92199b0%29> at
perf.c:357
#16 0x0000000000437157 in handle_internal_command (argv=0x7ffcd92199b0, 
argc=4) at perf.c:419
#17 run_argv (argv=0x7ffcd9219720, argcp=0x7ffcd921972c) at perf.c:465
#18 main (argc=4, argv=0x7ffcd92199b0) at perf.c:609

             reply	other threads:[~2016-08-05 10:58 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-08-05 10:58 Wangnan (F) [this message]
2016-08-05 14:12 ` [BUG] perf: memory leak in perf top Arnaldo Carvalho de Melo
2016-08-08 11:34   ` Wangnan (F)
2016-08-08 13:29     ` 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=57A47144.6000406@huawei.com \
    --to=wangnan0@huawei.com \
    --cc=acme@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=namhyung@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.