linux-kernel - Re: [PATCH 0/4] perf: add support for profiling jitted code

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CABPqkBRDYg7vJaKzJmP0LDQZXkwtiY=SJ_jgoZiRy3yeFUbCWQ@mail.gmail.com>
Date:	Thu, 12 Feb 2015 12:27:01 -0500
From:	Stephane Eranian <eranian@...gle.com>
To:	Jiri Olsa <jolsa@...hat.com>
Cc:	LKML <linux-kernel@...r.kernel.org>,
	Arnaldo Carvalho de Melo <acme@...hat.com>,
	Peter Zijlstra <peterz@...radead.org>,
	"mingo@...e.hu" <mingo@...e.hu>,
	"ak@...ux.intel.com" <ak@...ux.intel.com>,
	Namhyung Kim <namhyung@...nel.org>,
	Rose Belcher <cel@...ibm.com>,
	Sukadev Bhattiprolu <sukadev@...ux.vnet.ibm.com>,
	Sonny Rao <sonnyrao@...omium.org>,
	John Mccutchan <johnmccutchan@...gle.com>
Subject: Re: [PATCH 0/4] perf: add support for profiling jitted code

Jiri,

On Thu, Feb 12, 2015 at 8:42 AM, Jiri Olsa <jolsa@...hat.com> wrote:
> On Wed, Feb 11, 2015 at 12:42:41AM +0100, Stephane Eranian wrote:
>
> SNIP
>
>> To use the new feature:
>>   - install the posix clock driver
>>   - make sure you chmod 644 /dev/trace_clock
>>   - compile perf
>>   - cd tools/perf/jvmti; make; install wherever is appropriate
>>
>>   Example using openJDK:
>>    $ perf record java -agentpath:libjvmti.so my_class
>>    java: jvmti: jitdump in $HOME/.debug/jit/java-jit-20150207.XXL9649H/jit-6320.dump
>>    $ perf inject -i perf.data -j $HOME/.debug/jit/java-jit-20150207.XXL9649H/jit-6320.dump -o perf.data.jitted
>>    $ perf report -i perf.data.jitted
>
> So report should display Java symbols now right? Is there something
> wrong with my test below?
>
That is because the timestamps are incorrect. Sorry about that.
Sonny's module as is is not going to generate the correct timestamps.
I will  V2 today. It uses another module from David and it works because
it is using the same time source as perf_events.

> thanks,
> jirka
>
>
> [jolsa@...-x3650m4-01 perf]$ ./perf record java -agentpath:./jvmti/libjvmti.so Puppy
> java: jvmti: jitdump in /home/jolsa/.debug/jit/java-jit-20150212.XXru95af/jit-3777.dump
> Passed Name is :tommy
> Puppy's age is :2
> Variable Value :2
> ^C[ perf record: Woken up 3 times to write data ]
> [ perf record: Captured and wrote 0.666 MB perf.data (17258 samples) ]
>
>
> [jolsa@...-x3650m4-01 perf]$ ./perf inject -v -i perf.data -j  /home/jolsa/.debug/jit/java-jit-20150212.XXru95af/jit-3777.dump -o perf.data.jitted
> build id event received for /lib/modules/3.19.0-rc7jit+/build/vmlinux: 4bb9dab963421f041b3eb3ec588fc07a58de4fa8
> build id event received for /usr/bin/bash: c9f090657c35c10d6edeca09f62de9d22060a706
> build id event received for /usr/lib64/ld-2.18.so: dddaf5704dbe9ace5c31f12ab90d8bfeb8fc5eb3
> build id event received for [vdso]: b9e5cb1dcb2ab0d9d969fdc0816c4ad9af37e26b
> build id event received for /usr/lib64/libc-2.18.so: f1b808c85f949e82cf864c1ef5a13c8acc703e3b
> build id event received for /usr/lib64/libpthread-2.18.so: 5b1e8bb0e7ae081585497d797e06614815b768ef
> build id event received for /usr/lib/jvm/java-1.8.0-openjdk-1.8.0.31.x86_64/jre/lib/amd64/server/libjvm.so: 7b87f33ae1b3150058839496dbe5b5521f8977b3
> build id event received for /usr/lib/jvm/java-1.8.0-openjdk-1.8.0.31.x86_64/jre/lib/amd64/libjava.so: 32594e2184996cd7ff8257c8ef70656d91d2b03c
> build id event received for /usr/lib/jvm/java-1.8.0-openjdk-1.8.0.31.x86_64/jre/lib/amd64/libzip.so: 5cbb2075b9b316fc39dd0d5f9b0dcc8825256a4b
> version=1
> size=32(0)
> ts=0x26feca
> pid=3777
> elf_mach=62
> failed to write feature 2
>
>
> [jolsa@...-x3650m4-01 perf]$ ./perf report -i perf.data.jitted --stdio | head -13
> Failed to open /tmp/perf-3777.map, continuing without symbols
> # To display the perf.data header info, please use --header/--header-only options.
> #
> # Samples: 17K of event 'cycles'
> # Event count (approx.): 9728779054
> #
> # Overhead  Command    Shared Object       Symbol
> # ........  .........  ..................  .....................................................
> #
>     50.90%  java-abrt  perf-3777.map       [.] 0x00007f93a501e550
>     15.00%  java-abrt  perf-3777.map       [.] 0x00007f93a501e586
>     14.16%  java-abrt  perf-3777.map       [.] 0x00007f93a501e579
>     12.63%  java-abrt  perf-3777.map       [.] 0x00007f93a501e67b
>      3.04%  java-abrt  perf-3777.map       [.] 0x00007f93a501e685
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/