powertop: powertop.c:113: push_line_pid: Assertion `strlen(string) > 0' failed.

[Zaki Manian]

Binary package hint: powertop

Powertop is routinely crashing on my Core i7 machine.

This message before crash

powertop: powertop.c:113: push_line_pid: Assertion `strlen(string) > 0' failed.
                                                                               Aborted (core dumped)

ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: powertop 1.11+git20091208-0ubuntu1
ProcVersionSignature: Ubuntu 2.6.32-18.27-generic 2.6.32.10+drm33.1
Uname: Linux 2.6.32-18-generic x86_64
NonfreeKernelModules: nvidia
Architecture: amd64
Date: Sat Apr 3 23:16:16 2010
EcryptfsInUse: Yes
InstallationMedia: Error: [Errno 13] Permission denied: '/var/log/installer/media-info'
ProcEnviron:
 LANG=en_US.utf8
 SHELL=/bin/bash
SourcePackage: powertop

[Данило Шеган (danilo)]

I usually see this with powertop when I plug the power cable in on a Lucid installation on Sony Vaio Z790 (core 2 duo) as wel.

[Anthony Batchelor (toeknee)]

For me, it usually happens when powertop is in the background. I've also had it happen when opening a new tab in terminator,

[ischoegl (ischoegl)]

'sudo powertop' crashes on lucid 64bit (both with/without power cord attached). The system is a thinkpad T410 (intel i5). No crash if run without root privileges.

[Nicolás Wolovick (nwolovick)]

I could not reproduce the situation that triggers the assert.
I have cloned the git repo

git clone git://git.moblin.org/powertop

And it is the same, always in the root account.
I have generated a coredump, but optimizations prevent to read some values.
Extrangely enough, putting in the Makefile no optimizations, seems to avoid such random failing assertion.

CFLAGS?=-O0 -g ${WARNFLAGS}

Does anyone can confirm this behaviour?

[Nicolás Wolovick (nwolovick)]

I have the following pattern.

-O0, -O3: no failing assert
-O1, -O2: failing assert

But no apparent reason since -O3 is a superset of -O2

gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
diff /tmp/O2-opts /tmp/O3-opts | grep enabled

> -fgcse-after-reload [enabled]
> -finline-functions [enabled]
> -fipa-cp-clone [enabled]
> -fpredictive-commoning [enabled]
> -ftree-vectorize [enabled]
> -funswitch-loops [enabled]

[Nicolás Wolovick (nwolovick)]

Here is a coredump and some analysis with gdb

Core was generated by `./powertop'.
Program terminated with signal 6, Aborted.
#0 0x00007f7f1f7f7a75 in *__GI_raise (sig=<value optimized out>) at ../nptl/sysdeps/unix/sysv/linux/raise.c:64
64 ../nptl/sysdeps/unix/sysv/linux/raise.c: No such file or directory.
 in ../nptl/sysdeps/unix/sysv/linux/raise.c
(gdb) bt
#0 0x00007f7f1f7f7a75 in *__GI_raise (sig=<value optimized out>) at ../nptl/sysdeps/unix/sysv/linux/raise.c:64
#1 0x00007f7f1f7fb5c0 in *__GI_abort () at abort.c:92
#2 0x00007f7f1f7f0941 in *__GI___assert_fail (assertion=0x40c111 "strlen(string) > 0", file=<value optimized out>, line=113,
    function=0x40d5a8 "push_line_pid") at assert.c:81
#3 0x0000000000402b89 in push_line_pid (string=0x7fffdea8df80 "", cpu_count=<value optimized out>, disk_count=<value optimized out>,
    pid=0x7fffdea8ff80 "3552") at powertop.c:113
#4 0x000000000040a953 in parse_event (ptr=<value optimized out>, verbose=0) at perf.c:230
#5 0x000000000040aaf6 in parse_data_dirty_buffer () at perf.c:270
#6 0x0000000000405035 in main (argc=<value optimized out>, argv=<value optimized out>) at powertop.c:1089
(gdb) up
#1 0x00007f7f1f7fb5c0 in *__GI_abort () at abort.c:92
92 abort.c: No such file or directory.
 in abort.c
(gdb) up
#2 0x00007f7f1f7f0941 in *__GI___assert_fail (assertion=0x40c111 "strlen(string) > 0", file=<value optimized out>, line=113,
    function=0x40d5a8 "push_line_pid") at assert.c:81
81 assert.c: No such file or directory.
 in assert.c
(gdb) up
#3 0x0000000000402b89 in push_line_pid (string=0x7fffdea8df80 "", cpu_count=<value optimized out>, disk_count=<value optimized out>,
    pid=0x7fffdea8ff80 "3552") at powertop.c:113
113 assert(strlen(string) > 0);
(gdb) up
#4 0x000000000040a953 in parse_event (ptr=<value optimized out>, verbose=0) at perf.c:230
230 push_line_pid(line, 0, 1, pid);
(gdb) print line
$1 = "\000hromium-browse\000omium-browse' is writing to file 'History-journal' on /dev/sda7.\nThis prevents the disk from going to powersave mode.\000\000\000\227\004\201\037\177\177\000\000\260\342\250\336\377\177\000\000\243\005\201\037\177\177\000\000\002\000\000\000\000\000\000\000\070\000\000\000\061\000\000\000.\000\000\000\070\000\000\000\201\v\201\037\177\177\000\000 113909 10502746\000\000\000\000\000\000\000\240"...
(gdb)

[Alex Bennée (ajbennee)]

The core dump is because of the assertion failure. Somehow the processing of /proc/timer_stats has broken. I can't see anything obvious as to why. I'd suggest running it under valgrind and seeing if that throws up any warnings.

I have seen some other "interesting" core dumps on 10.04 in various libc string functions in other programs so I wouldn't rule out a library problem.

[Kristofer (kbitner)]

I have this issue with Lucid 64 bit on a lenovo T410 with core i7. I execute 'sudo powertop' and a couple of seconds later the only thing on the screen is the mouse cursor and the computer freezes.

[Hasenkopf, Andreas (andi-hasenkopf)]

Some have experienced this problem under 10.04, but not me. On my machine (Thinkpad X200, Intel(R) Core(TM)2 Duo CPU P8400 @ 2.26GHz) powertop was working fine on 10.04.
After installation of 11.04 powertop goes crazy on my machine, too.