MATE power statistics crashed

[Martin Wimpress]

MATE Power Statistics will crash. Just click between a few tabs and power sources and it will soon die :(

ProblemType: Crash
DistroRelease: Ubuntu 15.04
Package: mate-power-manager 1.8.1+dfsg1-3
ProcVersionSignature: Ubuntu 3.19.0-7.7-generic 3.19.0
Uname: Linux 3.19.0-7-generic x86_64
ApportVersion: 2.16.2-0ubuntu1
Architecture: amd64
CurrentDesktop: MATE
Date: Wed Mar 4 21:26:28 2015
ExecutablePath: /usr/bin/mate-power-statistics
InstallationDate: Installed on 2015-03-04 (0 days ago)
InstallationMedia: Ubuntu-MATE 15.04 "Vivid Vervet" - Alpha amd64 (20150304)
ProcCmdline: /usr/bin/mate-power-statistics --device /org/freedesktop/UPower/devices/battery_BAT0
ProcEnviron:
 LANGUAGE=en_GB:en
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=en_GB.UTF-8
 SHELL=/bin/bash
SegvAnalysis: Skipped: missing required field "Disassembly"
Signal: 11
SourcePackage: mate-power-manager
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm cdrom dip lpadmin plugdev sambashare sudo

[Apport retracing service (apport)]

StacktraceTop:
 g_variant_unref (value=0x3) at /build/buildd/glib2.0-2.43.90/./glib/gvariant-core.c:625
 up_device_get_statistics_sync (device=0x7f7734017770, type=0x40e146 "discharging", cancellable=0x7f7743fa12f9 <up_device_get_statistics_sync+313>, cancellable@entry=0x0, error=0x0) at up-device.c:560
 gpm_stats_update_info_page_stats (device=0x7f7734017770) at gpm-statistics.c:708
 gpm_stats_update_info_data_page (device=device@entry=0x7f7734017770, page=<optimized out>) at gpm-statistics.c:759
 gpm_stats_update_info_data (device=device@entry=0x7f7734017770) at gpm-statistics.c:806

[Martin Wimpress (flexiondotorg)]

The following snippet is the patch confirmed by Fedora as fixing this issue.

  * http://paste.fedoraproject.org/192697/42539076/

[Martin Wimpress (flexiondotorg)]

The bug is actually in upower.

[Jason Gerecke (killertofu)]

Created attachment 114095
Patch to fix issue

If the command `upower -d` is run while a device with no history is attached, the command can crash with a segfault.

Relevant backtrace when compiled from the master branch (f3706d7) with debugging:

Program received signal SIGSEGV, Segmentation fault.
0x00007ffff76d8444 in g_variant_unref () from /usr/lib/libglib-2.0.so.0
(gdb) bt
#0 0x00007ffff76d8444 in g_variant_unref () from /usr/lib/libglib-2.0.so.0
#1 0x00007ffff7bc27b2 in up_device_get_history_sync (device=0x7ffff0003c70,
    type=0x7ffff7bd23af "charge", timespec=120, resolution=10, cancellable=0x0,
    error=0x0) at up-device.c:484
#2 0x00007ffff7bc192d in up_device_to_text_history (device=0x7ffff0003c70,
    string=0x7ffff0001920, type=0x7ffff7bd23af "charge") at up-device.c:194
#3 0x00007ffff7bc2391 in up_device_to_text (device=0x7ffff0003c70)
    at up-device.c:375
#4 0x0000000000401d14 in main (argc=1, argv=0x7fffffffe748) at up-tool.c:321
(gdb)

This appears to be caused by the fact that if the call to 'up_device_glue_call_get_history_sync' within the 'up_device_get_history_sync' function fails, the 'gva' variable will be left uninitialized. The function later tries to call 'g_variant_unref' on it and fails.

A patch which initializes 'gva' to NULL is attached and resolves the issue for me. A similar fix may also be needed for the 'up_device_get_statistics_sync' function (which follows the same pattern of leaving 'gva' uninitialized).

[Martin Pitt (pitti)]

*** Bug 89514 has been marked as a duplicate of this bug. ***

[Martin Pitt (pitti)]

I came here from a downstream report/patch, so I didn't take your patch, sorry.

Fixed in master now:

http://cgit.freedesktop.org/upower/commit/?id=2510148b16

Thanks!

[Martin Pitt (pitti)]

Fixed upstream now, will upload to Debian and sync ASAP.

[Launchpad Janitor (janitor)]

This bug was fixed in the package upower - 0.99.2-2

---------------
upower (0.99.2-2) experimental; urgency=medium

  * libupower: Fix crash due to uninitialized GVariant. (Closes: #774546,
    LP: #1428337)

 -- Martin Pitt <email address hidden> Fri, 13 Mar 2015 15:59:59 +0100