lkml.org 
[lkml]   [2010]   [May]   [6]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Tracing of power:power_start events doesn't work
On Wed 2010-05-05 19:23, Frederic Weisbecker wrote:
> On Wed, May 05, 2010 at 08:34:17AM +0200, Ronny Tschüter wrote:
>
>> On Wed 2010-05-05 02:31, Steven Rostedt wrote:
>>
>>> On Mon, 2010-05-03 at 15:19 +0200, Ronny Tschüter wrote:
>>>
>>>
>>>> Hello,
>>>>
>>>> i use kernel version 2.6.33.2 (x86-64, Core 2 Duo) and the perf events
>>>> to track frequency and c states of my cpu. But if i want to trace
>>>> power:power_start events, it does not work. After typing
>>>>
>>>> echo "power:power_start" | sudo tee set_event
>>>>
>>>> on the console, the trace doesn't contain such events.
>>>> power:power_frequency and power:power_end events are reported correctly
>>>> by the system. I read in the mailing list that Robert Schoene had a
>>>> similiar problem, but with power:power_end events. Is there a known bug
>>>> regarding to trace_power_start methods in process.c?
>>>>
>>>>
>>> Is this for perf or ftrace? I'm assuming you did this from the debugfs
>>> tracing directory. The set_event file is for ftrace not perf.
>>>
>>>
>> It is for perf - i'm interested in event tracing. I used the event
>> tracing system like it is described in Documentation/trace/events.txt.
>> In addition i wrote a program which uses syscall(__NR_perf_counter_open,
>> ... ) to trace power events. But the result is the same -
>> power:power_end events are reported, power:power_start events not.
>>
>> Bye Ronny
>>
>
> Hi, can you please compare the perf and ftrace output to
> see if ftrace works while perf doesn't for example?
>
> In perf:
> perf record -f -c 1 -a -e power:power_start
> ^C
> perf report
>
> In ftrace:
> echo 1> $DEBUGFS/tracing/events/power/power_start/enable
> cat $DEBUGFS/trace_pipe
>
> Thanks
Hi

IN PERF:

-> patched kernel (see my mail from yesterday)
perf record -f -c 1 -a -e power:power_start
^C[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.500 MB perf.data (~21826 samples) ]
perf report
# Samples: 7880
#
# Overhead Command Shared Object Symbol
# ........ ............... ................. ......
#
57.13% swapper 0 [k] 0000000000000000
42.86% init 0 [k] 0000000000000000
0.01% swapper [kernel] [k] 0000000000000000
#
# (For a higher level overview, try: perf report --sort comm,dso)
#

-> unpatched kernel
perf record -f -c 1 -a -e power:power_start
^C[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.263 MB perf.data (~11484 samples) ]
perf report
# Samples: 0
#
# Overhead Command Shared Object Symbol
# ........ ............... ............. ......
#
#
# (For a higher level overview, try: perf report --sort comm,dso)
#


IN FTRACE:
-> patched kernel
cat trace_pipe shows events like
<idle>-0 [001] 59067.030622: power_start: type=1 state=2
<idle>-0 [001] 59067.032514: power_start: type=1 state=2
<idle>-0 [000] 59067.032522: power_start: type=1 state=2
<idle>-0 [001] 59067.032561: power_start: type=1 state=2
<idle>-0 [000] 59067.037527: power_start: type=1 state=2
<idle>-0 [001] 59067.037528: power_start: type=1 state=2

-> unpatched kernel
cat trace returns no output

Bye
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2010-05-06 08:35    [W:0.070 / U:2.628 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site