4:25 a.m.
Hi Marc and I were debugging segfault in Clear Linux version 28020
Program received signal SIGSEGV, Segmentation fault.
0x0000000000442b66
in perf_event::process (this=0x4a3ab0, cookie=cookie@entry=0x4b1d10)
at perf/perf.cpp:233
233 if (perf_fd < 0)
(gdb) bt
#0 0x0000000000442b66 in perf_event::process (this=0x4a3ab0,
cookie=cookie@entry=0x4b1d10) at perf/perf.cpp:233
#1 0x00000000004439d8 in perf_bundle::process (this=0x4b1cc0) at
perf/perf_bundle.cpp:326
#2 0x0000000000420ce1 in process_cpu_data () at cpu/cpu.cpp:1037
#3 0x00000000004146d5 in one_measurement (seconds=-4,
seconds@entry=1, sample_interval=<optimized out>,
sample_interval@entry=5, workload=workload@entry=0x0) at main.cpp:255
#4 0x000000000040aad3 in main (argc=1, argv=0x7fffffffcf58) at main.cpp:542
The bug seems to be in this call of the perf file descriptor
https://github.com/fenrus75/powertop/blob/master/src/perf/perf_bundle.cpp...
by the call of perf_event::process in perf.cpp
I guess that perf_fd = sys_perf_event_open(&attr, -1, _cpu, -1, 0); is
not generating the perf event
We have a copy of perf_event.h taken from the kernel and I wonder if
we do need to update it
In the Clear Linux we are running kernel: 4.20.10-698.native
Happy to help with the fix, just need a bit of direction
Regards
Victor Rodriguez
8:05 a.m.
New subject: [Powertop] seg fault at clearlinux v2.10 with kernel 4.20
Hi Victor,
4.20.10 has a kernel bug that breaks powertop. Please refer this
https://lists.01.org/pipermail/powertop/2019-February/002040.html . Also, try latest Clear
Linux version 28100.
Thanks,
Gayatri
-----Original Message-----
From: PowerTop [mailto:powertop-bounces@lists.01.org] On Behalf Of
Victor Rodriguez
Sent: Sunday, March 3, 2019 9:25 AM
To: powertop(a)lists.01.org; Simental Magana, Marcos
<marcos.simental.magana(a)intel.com>
Subject: [Powertop] seg fault at clearlinux v2.10 with kernel 4.20
Hi Marc and I were debugging segfault in Clear Linux version 28020
Program received signal SIGSEGV, Segmentation fault.
0x0000000000442b66 in
perf_event::process
(this=0x4a3ab0, cookie=cookie@entry=0x4b1d10) at perf/perf.cpp:233
233 if (perf_fd < 0)
(gdb) bt
#0 0x0000000000442b66 in perf_event::process (this=0x4a3ab0,
cookie=cookie@entry=0x4b1d10) at perf/perf.cpp:233
#1 0x00000000004439d8 in perf_bundle::process (this=0x4b1cc0) at
perf/perf_bundle.cpp:326
#2 0x0000000000420ce1 in process_cpu_data () at cpu/cpu.cpp:1037
#3 0x00000000004146d5 in one_measurement (seconds=-4,
seconds@entry=1, sample_interval=<optimized out>,
sample_interval@entry=5, workload=workload@entry=0x0) at
main.cpp:255
#4 0x000000000040aad3 in main (argc=1, argv=0x7fffffffcf58) at
main.cpp:542
The bug seems to be in this call of the perf file descriptor
https://github.com/fenrus75/powertop/blob/master/src/perf/perf_bundle.c
pp#L323
by the call of perf_event::process in perf.cpp
I guess that perf_fd = sys_perf_event_open(&attr, -1, _cpu, -1, 0); is not
generating the perf event
We have a copy of perf_event.h taken from the kernel and I wonder if we do
need to update it
In the Clear Linux we are running kernel: 4.20.10-698.native
Happy to help with the fix, just need a bit of direction
Regards
Victor Rodriguez
_______________________________________________
PowerTop mailing list
PowerTop(a)lists.01.org
https://lists.01.org/mailman/listinfo/powertop
12:51 a.m.
New subject: [Powertop] seg fault at clearlinux v2.10 with kernel 4.20
Fixed , thanks :)
On Sun, Mar 3, 2019 at 3:05 PM Kammela, Gayatri
<gayatri.kammela(a)intel.com> wrote:
Hi Victor,
4.20.10 has a kernel bug that breaks powertop. Please refer this
https://lists.01.org/pipermail/powertop/2019-February/002040.html . Also, try latest Clear
Linux version 28100.
Thanks,
Gayatri
> -----Original Message-----
> From: PowerTop [mailto:powertop-bounces@lists.01.org] On Behalf Of
> Victor Rodriguez
> Sent: Sunday, March 3, 2019 9:25 AM
> To: powertop(a)lists.01.org; Simental Magana, Marcos
> <marcos.simental.magana(a)intel.com>
> Subject: [Powertop] seg fault at clearlinux v2.10 with kernel 4.20
>
> Hi Marc and I were debugging segfault in Clear Linux version 28020
>
> Program received signal SIGSEGV, Segmentation fault.
> 0x0000000000442b66 in
perf_event::process
> (this=0x4a3ab0, cookie=cookie@entry=0x4b1d10) at perf/perf.cpp:233
> 233 if (perf_fd < 0)
> (gdb) bt
> #0 0x0000000000442b66 in perf_event::process (this=0x4a3ab0,
> cookie=cookie@entry=0x4b1d10) at perf/perf.cpp:233
> #1 0x00000000004439d8 in perf_bundle::process (this=0x4b1cc0) at
> perf/perf_bundle.cpp:326
> #2 0x0000000000420ce1 in process_cpu_data () at cpu/cpu.cpp:1037
> #3 0x00000000004146d5 in one_measurement (seconds=-4,
> seconds@entry=1, sample_interval=<optimized out>,
> sample_interval@entry=5, workload=workload@entry=0x0) at
> main.cpp:255
> #4 0x000000000040aad3 in main (argc=1, argv=0x7fffffffcf58) at
> main.cpp:542
>
> The bug seems to be in this call of the perf file descriptor
> https://github.com/fenrus75/powertop/blob/master/src/perf/perf_bundle.c
> pp#L323
>
> by the call of perf_event::process in perf.cpp
>
> I guess that perf_fd = sys_perf_event_open(&attr, -1, _cpu, -1, 0); is not
> generating the perf event
>
> We have a copy of perf_event.h taken from the kernel and I wonder if we do
> need to update it
>
> In the Clear Linux we are running kernel: 4.20.10-698.native
>
> Happy to help with the fix, just need a bit of direction
>
> Regards
>
> Victor Rodriguez
> _______________________________________________
> PowerTop mailing list
> PowerTop(a)lists.01.org
> https://lists.01.org/mailman/listinfo/powertop
1294
days inactive
1295
days old
2 comments
2 participants
participants (2)
-
Kammela, Gayatri -
Victor Rodriguez