BUG: perf top enters loop synthesizing events for existing threads

public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed

* BUG: perf top enters loop synthesizing events for existing threads
@ 2015-04-08 13:42 Arnaldo Carvalho de Melo
  2015-04-08 13:45 ` Arnaldo Carvalho de Melo
  0 siblings, 1 reply; 6+ messages in thread
From: Arnaldo Carvalho de Melo @ 2015-04-08 13:42 UTC (permalink / raw)
  To: David Ahern
  Cc: Linux Kernel Mailing List, Don Zickus, Jiri Olsa, Joe Mario,
	Ingo Molnar

Heads up,

	While testing some unrelated patches I noticed, in one of my
machines, that 'perf top', as root, is entering some sort of loop and
ends up consuming a lot of memory, making the system unresponsive,
bisected it down to:

[acme@zoo linux]$ git bisect good
ca6c41c59b964d362823e80442e9e32c31106b29 is the first bad commit
commit ca6c41c59b964d362823e80442e9e32c31106b29
Author: David Ahern <dsahern@gmail.com>
Date:   Mon Mar 30 14:35:58 2015 -0600

    perf tools: Fix ppid for synthesized fork events
    
    363b785f38 added synthesized fork events and set a thread's parent id to
    itself. Since we are already processing /proc/<pid>/status the ppid can
    be determined properly. Make it so.
    
    Signed-off-by: David Ahern <dsahern@gmail.com>
    Acked-by: Don Zickus <dzickus@redhat.com>
    Acked-by: Jiri Olsa <jolsa@kernel.org>
    Cc: Joe Mario <jmario@redhat.com>
    Link: http://lkml.kernel.org/r/1427747758-18510-2-git-send-email-dsahern@gmail.com
    Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

:040000 040000 24d479e4afbe95aaba882850473a870b6b3c7e87
cc86fe0e51b5e8a66e8a23f934fd0c4ff0f8dcd6 M	tools
[acme@zoo linux]$


I also noticed that if I set /proc/sys/kernel/perf_event_paranoid to -1
to run it as !root, then the problem "goes away", which I think probably
is explained by, as !root, not being able to parse some of the /proc
files for existing threads and thus not triggering the bug, still
investigating...

I might need to revert it for this cycle...

- Arnaldo

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: BUG: perf top enters loop synthesizing events for existing threads
  2015-04-08 13:42 BUG: perf top enters loop synthesizing events for existing threads Arnaldo Carvalho de Melo
@ 2015-04-08 13:45 ` Arnaldo Carvalho de Melo
  2015-04-08 13:48   ` David Ahern
  0 siblings, 1 reply; 6+ messages in thread
From: Arnaldo Carvalho de Melo @ 2015-04-08 13:45 UTC (permalink / raw)
  To: David Ahern
  Cc: Linux Kernel Mailing List, Don Zickus, Jiri Olsa, Joe Mario,
	Ingo Molnar

Em Wed, Apr 08, 2015 at 10:42:51AM -0300, Arnaldo Carvalho de Melo escreveu:
> Heads up,
> 
> 	While testing some unrelated patches I noticed, in one of my
> machines, that 'perf top', as root, is entering some sort of loop and
> ends up consuming a lot of memory, making the system unresponsive,
> bisected it down to:
> 
> [acme@zoo linux]$ git bisect good
> ca6c41c59b964d362823e80442e9e32c31106b29 is the first bad commit
> commit ca6c41c59b964d362823e80442e9e32c31106b29
> Author: David Ahern <dsahern@gmail.com>
> Date:   Mon Mar 30 14:35:58 2015 -0600
> 
>     perf tools: Fix ppid for synthesized fork events
>     
>     363b785f38 added synthesized fork events and set a thread's parent id to
>     itself. Since we are already processing /proc/<pid>/status the ppid can
>     be determined properly. Make it so.
>     
>     Signed-off-by: David Ahern <dsahern@gmail.com>
>     Acked-by: Don Zickus <dzickus@redhat.com>
>     Acked-by: Jiri Olsa <jolsa@kernel.org>
>     Cc: Joe Mario <jmario@redhat.com>
>     Link: http://lkml.kernel.org/r/1427747758-18510-2-git-send-email-dsahern@gmail.com
>     Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
> 
> :040000 040000 24d479e4afbe95aaba882850473a870b6b3c7e87
> cc86fe0e51b5e8a66e8a23f934fd0c4ff0f8dcd6 M	tools
> [acme@zoo linux]$
> 
> 
> I also noticed that if I set /proc/sys/kernel/perf_event_paranoid to -1
> to run it as !root, then the problem "goes away", which I think probably
> is explained by, as !root, not being able to parse some of the /proc
> files for existing threads and thus not triggering the bug, still
> investigating...
> 
> I might need to revert it for this cycle...

Just reverting it makes the symptoms go away, to works as before. Trying
to understand the problem now.

- Arnaldo

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: BUG: perf top enters loop synthesizing events for existing threads
  2015-04-08 13:45 ` Arnaldo Carvalho de Melo
@ 2015-04-08 13:48   ` David Ahern
  2015-04-08 14:14     ` Arnaldo Carvalho de Melo
  0 siblings, 1 reply; 6+ messages in thread
From: David Ahern @ 2015-04-08 13:48 UTC (permalink / raw)
  To: Arnaldo Carvalho de Melo
  Cc: Linux Kernel Mailing List, Don Zickus, Jiri Olsa, Joe Mario,
	Ingo Molnar

On 4/8/15 7:45 AM, Arnaldo Carvalho de Melo wrote:
>> I also noticed that if I set /proc/sys/kernel/perf_event_paranoid to -1
>> to run it as !root, then the problem "goes away", which I think probably
>> is explained by, as !root, not being able to parse some of the /proc
>> files for existing threads and thus not triggering the bug, still
>> investigating...

odd. because this has nothing to do with perf_events; it is just walking 
/proc and for the ppid adds a strstr and atoi(str). Something else is at 
play. I'll take a look.

David

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: BUG: perf top enters loop synthesizing events for existing threads
  2015-04-08 13:48   ` David Ahern
@ 2015-04-08 14:14     ` Arnaldo Carvalho de Melo
  2015-04-08 14:34       ` Arnaldo Carvalho de Melo
  0 siblings, 1 reply; 6+ messages in thread
From: Arnaldo Carvalho de Melo @ 2015-04-08 14:14 UTC (permalink / raw)
  To: David Ahern
  Cc: Linux Kernel Mailing List, Don Zickus, Jiri Olsa, Joe Mario,
	Ingo Molnar

Em Wed, Apr 08, 2015 at 07:48:23AM -0600, David Ahern escreveu:
> On 4/8/15 7:45 AM, Arnaldo Carvalho de Melo wrote:
> >>I also noticed that if I set /proc/sys/kernel/perf_event_paranoid to -1
> >>to run it as !root, then the problem "goes away", which I think probably
> >>is explained by, as !root, not being able to parse some of the /proc
> >>files for existing threads and thus not triggering the bug, still
> >>investigating...
> 
> odd. because this has nothing to do with perf_events; it is just walking
> /proc and for the ppid adds a strstr and atoi(str). Something else is at
> play. I'll take a look.

You are correcly setting the pid values, that will have effects when
using findnew, i.e. threads will be added to the rbtree, which causes
allocations, etc I.e. it is probably triggering a dormant bug

- Arnaldo

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: BUG: perf top enters loop synthesizing events for existing threads
  2015-04-08 14:14     ` Arnaldo Carvalho de Melo
@ 2015-04-08 14:34       ` Arnaldo Carvalho de Melo
  2015-04-08 14:37         ` Arnaldo Carvalho de Melo
  0 siblings, 1 reply; 6+ messages in thread
From: Arnaldo Carvalho de Melo @ 2015-04-08 14:34 UTC (permalink / raw)
  To: David Ahern
  Cc: Linux Kernel Mailing List, Don Zickus, Jiri Olsa, Joe Mario,
	Ingo Molnar

Em Wed, Apr 08, 2015 at 11:14:27AM -0300, Arnaldo Carvalho de Melo escreveu:
> Em Wed, Apr 08, 2015 at 07:48:23AM -0600, David Ahern escreveu:
> > On 4/8/15 7:45 AM, Arnaldo Carvalho de Melo wrote:
> > >>I also noticed that if I set /proc/sys/kernel/perf_event_paranoid to -1
> > >>to run it as !root, then the problem "goes away", which I think probably
> > >>is explained by, as !root, not being able to parse some of the /proc
> > >>files for existing threads and thus not triggering the bug, still
> > >>investigating...
> > 
> > odd. because this has nothing to do with perf_events; it is just walking
> > /proc and for the ppid adds a strstr and atoi(str). Something else is at
> > play. I'll take a look.
> 
> You are correcly setting the pid values, that will have effects when
> using findnew, i.e. threads will be added to the rbtree, which causes
> allocations, etc I.e. it is probably triggering a dormant bug

Sitting there, will stop chrome to see if the problem is triggered by it...

Program received signal SIGTERM, Terminated.
0x00000000004abc39 in rb_next (node=0x16202d20) at ../../lib/rbtree.c:457
457			while (node->rb_left)
(gdb) bt
#0  0x00000000004abc39 in rb_next (node=0x16202d20) at ../../lib/rbtree.c:457
#1  0x00000000004ccd27 in map_groups__fixup_overlappings (mg=0x12745740, map=0x16676210, fp=0x7ffff5a711c0 <_IO_2_1_stderr_>) at util/map.c:643
#2  0x00000000004d457d in thread__insert_map (thread=0x12745680, map=0x16676210) at util/thread.c:177
#3  0x00000000004c8de3 in machine__process_mmap2_event (machine=0xa28570, event=0xa59410, sample=0x870160 <synth_sample>) at util/machine.c:1204
#4  0x00000000004c92c3 in machine__process_event (machine=0xa28570, event=0xa59410, sample=0x870160 <synth_sample>) at util/machine.c:1329
#5  0x0000000000494a81 in perf_event__process (tool=0x0, event=0xa59410, sample=0x870160 <synth_sample>, machine=0xa28570) at util/event.c:778
#6  0x0000000000493606 in perf_event__synthesize_mmap_events (tool=0x0, event=0xa59410, pid=2513, tgid=2513, process=0x494a42 <perf_event__process>, machine=0xa28570, mmap_data=false) at util/event.c:302
#7  0x0000000000493b7b in __event__synthesize_thread (comm_event=0xa2a820, mmap_event=0xa59410, fork_event=0xa2a6a0, pid=2513, full=1, process=0x494a42 <perf_event__process>, tool=0x0, machine=0xa28570, mmap_data=false)
    at util/event.c:429
#8  0x0000000000493fe8 in perf_event__synthesize_threads (tool=0x0, process=0x494a42 <perf_event__process>, machine=0xa28570, mmap_data=false) at util/event.c:550
#9  0x00000000004ca82e in __machine__synthesize_threads (machine=0xa28570, tool=0x0, target=0x7fffffffb008, threads=0x94e490, process=0x494a42 <perf_event__process>, data_mmap=false) at util/machine.c:1781
#10 0x000000000043ae4c in machine__synthesize_threads (machine=0xa28570, target=0x7fffffffb008, threads=0x94e490, data_mmap=false) at util/machine.h:218
#11 0x000000000043d744 in __cmd_top (top=0x7fffffffaf80) at builtin-top.c:952
#12 0x000000000043ea4f in cmd_top (argc=0, argv=0x7fffffffe3d0, prefix=0x0) at builtin-top.c:1267
#13 0x0000000000489132 in run_builtin (p=0x86f128 <commands+264>, argc=1, argv=0x7fffffffe3d0) at perf.c:370
#14 0x0000000000489391 in handle_internal_command (argc=1, argv=0x7fffffffe3d0) at perf.c:429
#15 0x00000000004894dd in run_argv (argcp=0x7fffffffe22c, argv=0x7fffffffe220) at perf.c:473
#16 0x000000000048983e in main (argc=1, argv=0x7fffffffe3d0) at perf.c:588
(gdb) 

(gdb) fr 3
#3  0x00000000004c8de3 in machine__process_mmap2_event (machine=0xa28570, event=0xa59410, sample=0x870160 <synth_sample>) at util/machine.c:1204
1204		thread__insert_map(thread, map);
(gdb) print *thread
$1 = {{rb_node = {__rb_parent_color = 37177040, rb_right = 0x0, rb_left = 0x0}, node = {next = 0x23746d0, prev = 0x0}}, mg = 0x12745740, pid_ = 2513, tid = 2513, ppid = 2480, cpu = -1, refcnt = 2, shortname = "\000\000", 
  comm_set = true, dead = false, comm_list = {next = 0x16f948b0, prev = 0x12745710}, comm_len = 0, db_id = 0, priv = 0x0, ts = 0x0}
(gdb) print event->header
$4 = {type = 10, misc = 2, size = 104}
(gdb)
[acme@zoo linux]$ ps ax|grep 2513
 2513 ?        S      0:00 /opt/google/chrome/chrome --type=gpu-broker

[acme@zoo linux]$ grep xp /proc/2513/maps 
7fac6ac94000-7fac6ac9c000 r-xp 00000000 fd:01 1187628    /usr/lib64/libpciaccess.so.0.11.1
7fac6ae9e000-7fac6aeaa000 r-xp 00000000 fd:01 1184221    /usr/lib64/libdrm_radeon.so.1.0.1
7fac6b0ab000-7fac6b0b1000 r-xp 00000000 fd:01 1184218    /usr/lib64/libdrm_nouveau.so.2.0.0
7fac6b2b2000-7fac6b2d2000 r-xp 00000000 fd:01 1184215    /usr/lib64/libdrm_intel.so.1.0.0
7fac6b4d3000-7fac6b9ef000 r-xp 00000000 fd:01 1442035    /usr/lib64/dri/i965_dri.so
7fac6bc1d000-7fac6bc2d000 r-xp 00000000 fd:01 1192079    /usr/lib64/libudev.so.1.4.0
7fac73553000-7fac735ca000 r-xp 00000000 fd:01 1185534    /usr/lib64/libfreebl3.so
7fac737d0000-7fac737d3000 r-xp 00000000 fd:01 1187499    /usr/lib64/libkeyutils.so.1.5
7fac739d4000-7fac739e1000 r-xp 00000000 fd:01 1189618    /usr/lib64/libkrb5support.so.0.1
7fac73be2000-7fac73be7000 r-xp 00000000 fd:01 1188148    /usr/lib64/libXxf86vm.so.1.0.0
7fac73de8000-7fac73dff000 r-xp 00000000 fd:01 1188964    /usr/lib64/libxcb-glx.so.0.0.0
7fac74002000-7fac74004000 r-xp 00000000 fd:01 1188124    /usr/lib64/libXau.so.6.0.0
7fac74206000-7fac74211000 r-xp 00000000 fd:01 1179773    /usr/lib64/libdrm.so.2.4.0
7fac74412000-7fac74437000 r-xp 00000000 fd:01 1190780    /usr/lib64/libglapi.so.0.0.0
7fac7463c000-7fac7464a000 r-xp 00000000 fd:01 1188942    /usr/lib64/libwayland-server.so.0.1.0
7fac7484c000-7fac74857000 r-xp 00000000 fd:01 1188939    /usr/lib64/libwayland-client.so.0.1.0
7fac74a59000-7fac74a66000 r-xp 00000000 fd:01 1187278    /usr/lib64/libgbm.so.1.0.0
7fac74c67000-7fac74c6a000 r-xp 00000000 fd:01 1188970    /usr/lib64/libxcb-shape.so.0.0.0
7fac74e6b000-7fac74e71000 r-xp 00000000 fd:01 1188976    /usr/lib64/libxcb-xfixes.so.0.0.0
7fac75072000-7fac75076000 r-xp 00000000 fd:01 1188963    /usr/lib64/libxcb-dri2.so.0.0.0
7fac75277000-7fac75278000 r-xp 00000000 fd:01 1188121    /usr/lib64/libX11-xcb.so.1.0.0
7fac75479000-7fac75493000 r-xp 00000000 fd:01 1188433    /usr/lib64/libgraphite2.so.3.0.1
7fac75695000-7fac756b9000 r-xp 00000000 fd:01 1184295    /usr/lib64/liblzma.so.5.0.99
7fac758ba000-7fac7591f000 r-xp 00000000 fd:01 1190568    /usr/lib64/libpcre.so.1.2.1
7fac75b20000-7fac75b22000 r-xp 00000000 fd:01 1188135    /usr/lib64/libXinerama.so.1.0.0
7fac75d23000-7fac75d43000 r-xp 00000000 fd:01 1188169    /usr/lib64/libatk-1.0.so.0.21009.1
7fac75f46000-7fac75f4e000 r-xp 00000000 fd:01 1188257    /usr/lib64/libcrypt-2.18.so
7fac7617d000-7fac7618d000 r-xp 00000000 fd:01 1188177    /usr/lib64/libavahi-client.so.3.2.9
7fac7638e000-7fac7639a000 r-xp 00000000 fd:01 1188178    /usr/lib64/libavahi-common.so.3.5.3
7fac7659b000-7fac76758000 r-xp 00000000 fd:01 1180383    /usr/lib64/libcrypto.so.1.0.1e
7fac76983000-7fac769e5000 r-xp 00000000 fd:01 1185907    /usr/lib64/libssl.so.1.0.1e
7fac76bef000-7fac76bf2000 r-xp 00000000 fd:01 1185817    /usr/lib64/libcom_err.so.2.1
7fac76df3000-7fac76e25000 r-xp 00000000 fd:01 1187372    /usr/lib64/libk5crypto.so.3.1
7fac77028000-7fac770f8000 r-xp 00000000 fd:01 1187503    /usr/lib64/libkrb5.so.3.3
7fac77308000-7fac7734f000 r-xp 00000000 fd:01 1189531    /usr/lib64/libgssapi_krb5.so.2.2
7fac77552000-7fac775e2000 r-xp 00000000 fd:01 1190785    /usr/lib64/libGL.so.1.2.0
7fac777e6000-7fac77806000 r-xp 00000000 fd:01 1188983    /usr/lib64/libxcb.so.1.1.0
7fac77a07000-7fac77a10000 r-xp 00000000 fd:01 1188967    /usr/lib64/libxcb-render.so.0.0.0
7fac77c11000-7fac77c13000 r-xp 00000000 fd:01 1188971    /usr/lib64/libxcb-shm.so.0.0.0
7fac77e14000-7fac77e3c000 r-xp 00000000 fd:01 1190790    /usr/lib64/libEGL.so.1.0.0
7fac7803e000-7fac780e0000 r-xp 00000000 fd:01 1187639    /usr/lib64/libpixman-1.so.0.30.0
7fac782e8000-7fac7833c000 r-xp 00000000 fd:01 1187448    /usr/lib64/libharfbuzz.so.0.927.0
7fac7853e000-7fac7853f000 r-xp 00000000 fd:01 1188489    /usr/lib64/libgthread-2.0.so.0.3800.2
7fac78740000-7fac78754000 r-xp 00000000 fd:01 1187618    /usr/lib64/libpangoft2-1.0.so.0.3600.1
7fac78956000-7fac78988000 r-xp 00000000 fd:01 1184251    /usr/lib64/libpng16.so.16.6.0
7fac78b89000-7fac78b9f000 r-xp 00000000 fd:01 1189224    /usr/lib64/libresolv-2.18.so
7fac78da3000-7fac78dc4000 r-xp 00000000 fd:01 1187969    /usr/lib64/libselinux.so.1
7fac78fc7000-7fac78fdc000 r-xp 00000000 fd:01 1188999    /usr/lib64/libz.so.1.2.8
7fac791dd000-7fac79203000 r-xp 00000000 fd:01 1188266    /usr/lib64/libdbus-glib-1.so.2.2.2
7fac79405000-7fac79408000 r-xp 00000000 fd:01 1188412    /usr/lib64/libgmodule-2.0.so.0.3800.2
7fac79609000-7fac7960c000 r-xp 00000000 fd:01 1189435    /usr/lib64/libplds4.so
7fac7980d000-7fac79814000 r-xp 00000000 fd:01 1188331    /usr/lib64/libffi.so.6.0.1
7fac79a15000-7fac79bc9000 r-xp 00000000 fd:01 1187054    /usr/lib64/libc-2.18.so
7fac79dd3000-7fac79de8000 r-xp 00000000 fd:01 1183174    /usr/lib64/libgcc_s-4.8.3-20140911.so.1
7fac79fe9000-7fac7a0d2000 r-xp 00000000 fd:01 1183177    /usr/lib64/libstdc++.so.6.0.19
7fac7a2f1000-7fac7a2f3000 r-xp 00000000 fd:01 1190263    /usr/lib64/libXss.so.1.0.0
7fac7a4f5000-7fac7a518000 r-xp 00000000 fd:01 1187313    /usr/lib64/libgdk_pixbuf-2.0.so.0.3000.3
7fac7a719000-7fac7a7d5000 r-xp 00000000 fd:01 1185350    /usr/lib64/libgdk-x11-2.0.so.0.2400.27
7fac7a9da000-7fac7ae52000 r-xp 00000000 fd:01 1185352    /usr/lib64/libgtk-x11-2.0.so.0.2400.27
7fac7b060000-7fac7b0a6000 r-xp 00000000 fd:01 1185836    /usr/lib64/libdbus-1.so.3.7.12
7fac7b2a8000-7fac7b2c0000 r-xp 00000000 fd:01 1188678    /usr/lib64/libpthread-2.18.so
7fac7b4c5000-7fac7b539000 r-xp 00000000 fd:01 1185342    /usr/lib64/libcups.so.2
7fac7b742000-7fac7b769000 r-xp 00000000 fd:01 1188325    /usr/lib64/libexpat.so.1.6.0
7fac7b769000-7fac7b969000 ---p 00027000 fd:01 1188325    /usr/lib64/libexpat.so.1.6.0
7fac7b969000-7fac7b96b000 r--p 00027000 fd:01 1188325    /usr/lib64/libexpat.so.1.6.0
7fac7b96b000-7fac7b96c000 rw-p 00029000 fd:01 1188325    /usr/lib64/libexpat.so.1.6.0
7fac7b96c000-7fac7b975000 r-xp 00000000 fd:01 1188139    /usr/lib64/libXrandr.so.2.2.0
7fac7bb76000-7fac7bb7b000 r-xp 00000000 fd:01 1188142    /usr/lib64/libXtst.so.6.1.0
7fac7bd7c000-7fac7bd7e000 r-xp 00000000 fd:01 1188127    /usr/lib64/libXdamage.so.1.1.0
7fac7bf7f000-7fac7c084000 r-xp 00000000 fd:01 1188374    /usr/lib64/libm-2.18.so
7fac7c286000-7fac7c367000 r-xp 00000000 fd:01 1188164    /usr/lib64/libasound.so.2.0.0
7fac7c56f000-7fac7c571000 r-xp 00000000 fd:01 1188125    /usr/lib64/libXcomposite.so.1.0.0
7fac7c772000-7fac7c77b000 r-xp 00000000 fd:01 1188140    /usr/lib64/libXrender.so.1.3.0
7fac7c97c000-7fac7c981000 r-xp 00000000 fd:01 1188131    /usr/lib64/libXfixes.so.3.1.0
7fac7cb82000-7fac7cb93000 r-xp 00000000 fd:01 1188130    /usr/lib64/libXext.so.6.4.0
7fac7cd94000-7fac7cd9e000 r-xp 00000000 fd:01 1188126    /usr/lib64/libXcursor.so.1.0.2
7fac7cf9f000-7fac7cfae000 r-xp 00000000 fd:01 1183298    /usr/lib64/libXi.so.6.1.0
7fac7d1af000-7fac7d2e7000 r-xp 00000000 fd:01 1188122    /usr/lib64/libX11.so.6.3.0
7fac7d4ed000-7fac7d535000 r-xp 00000000 fd:01 1187616    /usr/lib64/libpango-1.0.so.0.3600.1
7fac7d738000-7fac7d859000 r-xp 00000000 fd:01 1187140    /usr/lib64/libcairo.so.2.11400.0
7fac7da5f000-7fac7da6b000 r-xp 00000000 fd:01 1187617    /usr/lib64/libpangocairo-1.0.so.0.3600.1
7fac7dc6c000-7fac7dd0b000 r-xp 00000000 fd:01 1187649    /usr/lib64/libfreetype.so.6.10.2
7fac7df11000-7fac7df4c000 r-xp 00000000 fd:01 1187277    /usr/lib64/libfontconfig.so.1.8.0
7fac7e14e000-7fac7e2ac000 r-xp 00000000 fd:01 1188399    /usr/lib64/libgio-2.0.so.0.3800.2
7fac7e4b3000-7fac7e4e3000 r-xp 00000000 fd:01 1188370    /usr/lib64/libgconf-2.so.4.1.5
7fac7e6e5000-7fac7e71f000 r-xp 00000000 fd:01 1189378    /usr/lib64/libnspr4.so
7fac7e923000-7fac7e927000 r-xp 00000000 fd:01 1189431    /usr/lib64/libplc4.so
7fac7eb28000-7fac7eb4c000 r-xp 00000000 fd:01 1192057    /usr/lib64/libsmime3.so
7fac7ed4f000-7fac7ed74000 r-xp 00000000 fd:01 1189436    /usr/lib64/libnssutil3.so
7fac7ef7b000-7fac7f099000 r-xp 00000000 fd:01 1191883    /usr/lib64/libnss3.so
7fac7f2a2000-7fac7f3cb000 r-xp 00000000 fd:01 1188407    /usr/lib64/libglib-2.0.so.0.3800.2
7fac7f5ce000-7fac7f61d000 r-xp 00000000 fd:01 1188425    /usr/lib64/libgobject-2.0.so.0.3800.2
7fac7f81f000-7fac7f822000 r-xp 00000000 fd:01 1188290    /usr/lib64/libdl-2.18.so
7fac7fa23000-7fac7fa2a000 r-xp 00000000 fd:01 1189264    /usr/lib64/librt-2.18.so
7fac7fc2b000-7fac7fc4b000 r-xp 00000000 fd:01 1187898    /usr/lib64/ld-2.18.so
7fac7fe48000-7fac7fe49000 rwxp 00000000 00:00 0 
7fac7fe4d000-7fac84fa0000 r-xp 00000000 fd:01 917553     /opt/google/chrome/chrome
7ffdd8dea000-7ffdd8dec000 r-xp 00000000 00:00 0          [vdso]
ffffffffff600000-ffffffffff601000 r-xp 00000000 00:00 0  [vsyscall]
[acme@zoo linux]$ 

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: BUG: perf top enters loop synthesizing events for existing threads
  2015-04-08 14:34       ` Arnaldo Carvalho de Melo
@ 2015-04-08 14:37         ` Arnaldo Carvalho de Melo
  0 siblings, 0 replies; 6+ messages in thread
From: Arnaldo Carvalho de Melo @ 2015-04-08 14:37 UTC (permalink / raw)
  To: David Ahern
  Cc: Linux Kernel Mailing List, Don Zickus, Jiri Olsa, Joe Mario,
	Ingo Molnar

Em Wed, Apr 08, 2015 at 11:34:37AM -0300, Arnaldo Carvalho de Melo escreveu:
> Em Wed, Apr 08, 2015 at 11:14:27AM -0300, Arnaldo Carvalho de Melo escreveu:
> > Em Wed, Apr 08, 2015 at 07:48:23AM -0600, David Ahern escreveu:
> > > On 4/8/15 7:45 AM, Arnaldo Carvalho de Melo wrote:
> > > >>I also noticed that if I set /proc/sys/kernel/perf_event_paranoid to -1
> > > >>to run it as !root, then the problem "goes away", which I think probably
> > > >>is explained by, as !root, not being able to parse some of the /proc
> > > >>files for existing threads and thus not triggering the bug, still
> > > >>investigating...
> > > 
> > > odd. because this has nothing to do with perf_events; it is just walking
> > > /proc and for the ppid adds a strstr and atoi(str). Something else is at
> > > play. I'll take a look.
> > 
> > You are correcly setting the pid values, that will have effects when
> > using findnew, i.e. threads will be added to the rbtree, which causes
> > allocations, etc I.e. it is probably triggering a dormant bug
> 
> Sitting there, will stop chrome to see if the problem is triggered by it...

Yeah, stopping chrome gets the system to a state that doesn't make it
trigger the bug, why it works when your patch is reverted still a
puzzle...

- Arnaldo

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2015-04-08 14:37 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-04-08 13:42 BUG: perf top enters loop synthesizing events for existing threads Arnaldo Carvalho de Melo
2015-04-08 13:45 ` Arnaldo Carvalho de Melo
2015-04-08 13:48   ` David Ahern
2015-04-08 14:14     ` Arnaldo Carvalho de Melo
2015-04-08 14:34       ` Arnaldo Carvalho de Melo
2015-04-08 14:37         ` Arnaldo Carvalho de Melo

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox