[Powertop] Segmentation fault occurs with --calibrate
Hi,
With the new Powertop 2.0, `powertop --calibrate` segfaults. Debugging output
is attached:
Starting program: /tmp/powertop-2.0/src/powertop --calibrate
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Starting PowerTOP power estimate calibration
Calibrating idle
Program received signal SIGSEGV, Segmentation fault.
perf_bundle::start (this=0x0) at perf/perf_bundle.cpp:180
180 for (i = 0; i < events.size(); i++) {
(gdb) bt
#0 perf_bundle::start (this=0x0) at perf/perf_bundle.cpp:180
#1 0x000000000040f9b0 in start_cpu_measurement () at cpu/cpu.cpp:297
#2 0x0000000000442655 in one_measurement (seconds=15) at main.cpp:182
#3 0x000000000043f4c2 in idle_calibration () at calibrate/calibrate.cpp:423
#4 calibrate () at calibrate/calibrate.cpp:466
#5 0x000000000040665f in main (argc=2, argv=0x7fffffffe718) at main.cpp:325
(gdb) up
#1 0x000000000040f9b0 in start_cpu_measurement () at cpu/cpu.cpp:297
297 perf_events->start()
(gdb) print perf_events
$5 = (perf_bundle *) 0x0
The order of processing seems flawed. Immediately during argument processing,
calibrate() is called. But start_cpu_measurement requires some initialization
which is done by enumerate_cpus() in the main() function after arguments
processing. You know your code better, so I'll gratefully wait for your
patches.
Thanks,
Peter