linux-kernel - Re: [PATCH v3]: perf record: enable arbitrary event names thru name= modifier

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: <80a37269-e54d-0992-712a-ee72555a04d4@linux.intel.com>
Date:   Mon, 4 Jun 2018 18:22:43 +0300
From:   Alexey Budankov <alexey.budankov@...ux.intel.com>
To:     Arnaldo Carvalho de Melo <acme@...nel.org>
Cc:     Jiri Olsa <jolsa@...hat.com>,
        Peter Zijlstra <peterz@...radead.org>,
        Ingo Molnar <mingo@...nel.org>,
        Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
        Namhyung Kim <namhyung@...nel.org>,
        Andi Kleen <ak@...ux.intel.com>,
        linux-kernel <linux-kernel@...r.kernel.org>,
        linux-perf-users@...r.kernel.org
Subject: Re: [PATCH v3]: perf record: enable arbitrary event names thru name=
 modifier

On 04.06.2018 17:58, Arnaldo Carvalho de Melo wrote:
> Em Mon, Jun 04, 2018 at 05:51:03PM +0300, Alexey Budankov escreveu:
>> Hi,
>> On 04.06.2018 17:23, Arnaldo Carvalho de Melo wrote:
>>> Em Mon, Jun 04, 2018 at 05:19:11PM +0300, Alexey Budankov escreveu:
>>>> Hi,
>>>>
>>>> On 04.06.2018 16:58, Arnaldo Carvalho de Melo wrote:
>>>>> Em Mon, Jun 04, 2018 at 09:56:02AM +0200, Jiri Olsa escreveu:
>>>>>> On Mon, Jun 04, 2018 at 09:50:56AM +0300, Alexey Budankov wrote:
>>>>>>> Enable complex event names containing [.:=,] symbols to be encoded into Perf 
>>>>>>> trace using name= modifier e.g. like this:
>>>>>
>>>>>>> perf record -e cpu/name=\'OFFCORE_RESPONSE:request=DEMAND_RFO:response=L3_HIT.SNOOP_HITM\',\
>>>>>>> 		period=0x3567e0,event=0x3c,cmask=0x1/Duk ./futex
>>>>>
>>>>>>> Below is how it looks like in the report output. Please note explicit escaped 
>>>>>>> quoting at cmdline string in the header so that thestring can be directly reused
>>>>>>> for another collection in shell:
>>>>>
>>>>> Applied, but there are other places where we show event names, such as:
>>>>>
>>>>> [root@...et ~]# perf record -e cpu/name=\'OFFCORE_RESPONSE:request=DEMAND_RFO:response=L3_HIT.SNOOP_HITM\',period=0x3567e0,event=0x3c,cmask=0x1/Duk
>>>>> ^C[ perf record: Woken up 1 times to write data ]
>>>>> [ perf record: Captured and wrote 1.599 MB perf.data (704 samples) ]
>>>>>
>>>>> [root@...et ~]# perf evlist
>>>>> OFFCORE_RESPONSE:request=DEMAND_RFO:response=L3_HIT.SNOOP_HITM
>>>>> [root@...et ~]# perf evlist -v
>>>>> OFFCORE_RESPONSE:request=DEMAND_RFO:response=L3_HIT.SNOOP_HITM: type: 4, size: 112, config: 0x100003c, { sample_period, sample_freq }: 3500000, sample_type: IP|TID|TIME|CPU, disabled: 1, inherit: 1, pinned: 1, exclude_hv: 1, mmap: 1, comm: 1, task: 1, sample_id_all: 1, mmap2: 1, comm_exec: 1
>>>>> [root@...et ~]# 
>>>>>
>>>>> That I used to check if the period, etc were correctly set, etc. Perhaps
>>>>> we should add that \'\' there as well?
>>>>
>>>> Like this?
>>>>
>>>> name: 'OFFCORE_RESPONSE:request=DEMAND_RFO:response=L3_HIT.SNOOP_HITM', type: 4, size: 112, config: 0x100003c, { sample_period, sample_freq }: 3500000, sample_type: IP|TID|TIME|CPU, disabled: 1, inherit: 1, pinned: 1, exclude_hv: 1, mmap: 1, comm: 1, task: 1, sample_id_all: 1, mmap2: 1, comm_exec: 1
>>>
>>> Nope, I mean with the same intent you had when showing escaped single
>>> quotes in the 'perf report' header, that users would copy'n'paste and
>>> get something that works on the command line?
>>
>> Well ok, if there are other cases aside perf report that tool shows full 
>> command line ready for reusing then it definitely makes sense to print 
> 
> Ok, lets leave it like you did, i.e. just when command lines appear and
> people are likely to copy'n'paste them to repeat it.
> 
>> event names quoted there. If such cases were pointed out then they also 
>> could be addressed along with the unit/regression testing mentioned above.
> 
> So lests stick to just the unit/regression testing, can you take a look
> at that?

Sure. Is it enough to run perf record with some quoted event name to make 
sure it runs successfully?

Thanks,
Alexey

> 
> - Arnaldo
>