Large problem memory leakage

oxydium · 2018-02-12 14:04:24

@Trilby
I do the manipulations as I understand them, hence my mistakes.
If I was very good in English, that would be an advantage, but it's not the case....

@seth
Okay, so when the "cinnamon --replace" process starts to take a lot of ram, I throw valgrind.
with this command??

valgrind --tool=massif cinnamon

or rather this command.?

valgrind --tool=massif cinnamon --replace

or both commands.?

I try to do my best to understand you, and to understand all the manipulations, and to answer you in English not too bad.

seth · 2018-02-12 14:30:04

You cannot attach valgrind to a running process (certainly not with massif and memcheck, callgrind allows you to remote-trigger the measure time, but we're not interested in a performance check)
So you will have to run cinnamon in valgrind all the time.

There's another caveat that will likely hit here: you will probably be told that something in js38 leaks, but valgrind doesn't usably debug javascript. We'd then know "yes, you can leak memory w/ javascript", but that'd be it :-(
You'd have to throw in something like https://github.com/Doist/JavaScript-memory-leak-checker and that is certainly something somebody familiar w/ the js code in cinnamon should do.

oxydium · 2018-02-12 17:48:16

ha all right, So it's getting more complicated to run tests.

I left my computer running for 3 hours and 30 minutes.
The "cinnamon --replace" process takes 590 MB.
I ran a Valgrind test and I got this.
I don't know if it's meaningful and useful.
Unfortunately, I don't know how to interpret the results

valgrind --tool=callgrind cinnamon --replace
==5939== Callgrind, a call-graph generating cache profiler
==5939== Copyright (C) 2002-2017, and GNU GPL'd, by Josef Weidendorfer et al.
==5939== Using Valgrind-3.13.0 and LibVEX; rerun with -h for copyright info
==5939== Command: cinnamon --replace
==5939== 
==5939== For interactive control, run 'callgrind_control -h'.
==5939== brk segment overflow in thread #1: can't grow to 0x4a32000
==5939== (see section Limitations in user manual)
==5939== NOTE: further instances of this message will not be shown
Cjs-Message: JS LOG: About to start Cinnamon
St-Message: cogl npot texture sizes SUPPORTED
Cjs-Message: JS LOG: Cinnamon started at Mon Feb 12 2018 18:32:23 GMT+0100 (CET)
==5963== 
==5963== Events    : Ir
==5963== Collected : 1666654843
==5963== 
==5963== I   refs:      1,666,654,843
==5965== 
==5965== Events    : Ir
==5965== Collected : 2068492403
==5965== 
==5965== I   refs:      2,068,492,403
==5970== 
==5970== Events    : Ir
==5970== Collected : 2072049864
==5970== 
==5970== I   refs:      2,072,049,864
Cjs-Message: JS LOG: network applet: Cannot find connection for active (or connection cannot be read)

(cinnamon:5939): St-CRITICAL **: _st_create_shadow_material: assertion 'src_texture != COGL_INVALID_HANDLE' failed

(cinnamon:5939): Gtk-WARNING **: Theme directory 22//places of theme Numix-White has no size field


(cinnamon:5939): Gtk-WARNING **: Theme directory 22/panel of theme Numix has no size field


(cinnamon:5939): Gtk-WARNING **: Theme directory 22@2x/panel of theme Numix has no size field


(cinnamon:5939): Gtk-WARNING **: Theme directory 24/panel of theme Numix has no size field


(cinnamon:5939): Gtk-WARNING **: Theme directory 24@2x/panel of theme Numix has no size field


(cinnamon:5939): St-CRITICAL **: _st_create_shadow_material: assertion 'src_texture != COGL_INVALID_HANDLE' failed

(cinnamon:5939): St-CRITICAL **: _st_create_shadow_material: assertion 'src_texture != COGL_INVALID_HANDLE' failed

(cinnamon:5939): St-CRITICAL **: _st_create_shadow_material: assertion 'src_texture != COGL_INVALID_HANDLE' failed

(cinnamon:5939): St-CRITICAL **: _st_create_shadow_material: assertion 'src_texture != COGL_INVALID_HANDLE' failed

(cinnamon:5939): St-CRITICAL **: _st_create_shadow_material: assertion 'src_texture != COGL_INVALID_HANDLE' failed

(cinnamon:5939): St-CRITICAL **: _st_create_shadow_material: assertion 'src_texture != COGL_INVALID_HANDLE' failed

(cinnamon:5939): St-CRITICAL **: _st_create_shadow_material: assertion 'src_texture != COGL_INVALID_HANDLE' failed
openGL version 4.2 detected (GL3 Cogl Driver)
MetaSyncRing disabled: couldn't find required GL extensions, or the minimum safe openGL version was not met
==5991== 
==5991== Events    : Ir
==5991== Collected : 4477282285
==5991== 
==5991== I   refs:      4,477,282,285
Cjs-Message: JS LOG: Invalid network device type, is 14
Cjs-Message: JS LOG: network applet: Found connection for active
==5939== 
==5939== Process terminating with default action of signal 11 (SIGSEGV): dumping core
==5939==  Access not within mapped region at address 0x64
==5939==    at 0x8232550: cogl_texture_get_width (in /usr/lib/libcogl.so.20.4.2)
==5939==    by 0x4ED1427: ??? (in /usr/lib/libmuffin.so.0.0.0)
==5939==    by 0x4ED170A: ??? (in /usr/lib/libmuffin.so.0.0.0)
==5939==    by 0x62A0C4A: g_datalist_clear (in /usr/lib/libglib-2.0.so.0.5400.3)
==5939==    by 0x6032926: ??? (in /usr/lib/libgobject-2.0.so.0.5400.3)
==5939==    by 0x6032F74: g_object_unref (in /usr/lib/libgobject-2.0.so.0.5400.3)
==5939==    by 0x6B9370F: ??? (in /usr/lib/libcjs.so.0.0.0)
==5939==    by 0x6B96AA7: ??? (in /usr/lib/libcjs.so.0.0.0)
==5939==    by 0xEF3CEAA: ??? (in /usr/lib/libmozjs-38.so)
==5939==    by 0xEF3DB77: ??? (in /usr/lib/libmozjs-38.so)
==5939==    by 0xEF3FD33: ??? (in /usr/lib/libmozjs-38.so)
==5939==    by 0xEF40708: ??? (in /usr/lib/libmozjs-38.so)
==5939==  If you believe this happened as a result of a stack
==5939==  overflow in your program's main thread (unlikely but
==5939==  possible), you can try to increase the size of the
==5939==  main thread stack using the --main-stacksize= flag.
==5939==  The main thread stack size used in this run was 8388608.
==5939== 
==5939== Events    : Ir
==5939== Collected : 6925987549
==5939== 
==5939== I   refs:      6,925,987,549
Segmentation error (core dumped)

after Valgrind's test, Cinnamon crashed.
And in dmesg there is this line

[12248.756937] cinnamon[980]: segfault at 64 ip 00007efec1913550 sp 00007ffe346f3748 error 4 in libcogl.so.20.4.2[7efec18bb000+b5000]

and with this command, I get this

journalctl -p err
-- Logs begin at Mon 2018-02-12 12:54:30 CET, end at Mon 2018-02-12 18:39:15 CET. --
févr. 12 12:54:30 archlinux kernel: ACPI Error: Needed type [Reference], found [Integer] ffff8ec3857c8360 (20170728/exresop-103)
févr. 12 12:54:30 archlinux kernel: ACPI Exception: AE_AML_OPERAND_TYPE, While resolving operands for [OpcodeName unavailable] (20170728/dswexec-
févr. 12 12:54:30 archlinux kernel: ACPI Error: Method parse/execution failed \_PR.CPU0._PDC, AE_AML_OPERAND_TYPE (20170728/psparse-550)
févr. 12 15:07:52 arch kernel: watchdog: watchdog0: watchdog did not stop!
-- Reboot --
févr. 12 15:07:58 archlinux kernel: ACPI Error: Needed type [Reference], found [Integer] ffff92af457c8f78 (20170728/exresop-103)
févr. 12 15:07:58 archlinux kernel: ACPI Exception: AE_AML_OPERAND_TYPE, While resolving operands for [OpcodeName unavailable] (20170728/dswexec-
févr. 12 15:07:58 archlinux kernel: ACPI Error: Method parse/execution failed \_PR.CPU0._PDC, AE_AML_OPERAND_TYPE (20170728/psparse-550)
févr. 12 15:25:56 arch cinnamon-screensaver-pam-helper[2830]: pam_unix(cinnamon-screensaver:auth): conversation failed
févr. 12 15:25:56 arch cinnamon-screensaver-pam-helper[2830]: pam_unix(cinnamon-screensaver:auth): auth could not identify password for [arch]
févr. 12 18:24:34 arch cinnamon-screensaver-pam-helper[5514]: pam_unix(cinnamon-screensaver:auth): conversation failed
févr. 12 18:24:34 arch cinnamon-screensaver-pam-helper[5514]: pam_unix(cinnamon-screensaver:auth): auth could not identify password for [arch]

Thanks

seth · 2018-02-12 17:58:03

Again: callgrind is *worthless* itr.

oxydium · 2018-02-12 18:00:52

What do I wear then???
I don't understand

Last edited by oxydium (2018-02-12 18:01:15)

seth · 2018-02-12 19:22:29

"--tool=massif"

oxydium · 2018-02-14 10:32:27

Does not work with valgrind --tool=massif..
Cinnamon crash at the launch of valgrind.....

Arch Linux

#26 2018-02-12 14:04:24

Re: Large problem memory leakage

#27 2018-02-12 14:30:04

Re: Large problem memory leakage

#28 2018-02-12 17:48:16

Re: Large problem memory leakage

#29 2018-02-12 17:58:03

Re: Large problem memory leakage

#30 2018-02-12 18:00:52

Re: Large problem memory leakage

#31 2018-02-12 19:22:29

Re: Large problem memory leakage

#32 2018-02-14 10:32:27

Re: Large problem memory leakage

Board footer