linux-kernel - Re: [PATCH 11/13] perf: add code to support PERF_SAMPLE_BRANCH

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]

Date:	Tue, 10 Jan 2012 16:43:29 +0100
From:	Stephane Eranian <eranian@...gle.com>
To:	Arun Sharma <asharma@...com>
Cc:	linux-kernel@...r.kernel.org, peterz@...radead.org, mingo@...e.hu,
	acme@...radead.org, robert.richter@....com, ming.m.lin@...el.com,
	andi@...stfloor.org, ravitillo@....gov, vweaver1@...s.utk.edu
Subject: Re: [PATCH 11/13] perf: add code to support PERF_SAMPLE_BRANCH_STACK (v3)

On Tue, Jan 10, 2012 at 2:25 AM, Arun Sharma <asharma@...com> wrote:
> On 1/9/12 8:49 AM, Stephane Eranian wrote:
>>
>> From: Roberto Agostino Vitillo<ravitillo@....gov>
>>
>> This patch adds:
>> - ability to parse samples with PERF_SAMPLE_BRANCH_STACK
>> - sort on branches
>> - build histograms on branches
>>
> [..]
>>
>>  static struct sort_dimension sort_dimensions[] = {
>> -       { .name = "pid",        .entry =&sort_thread,   },
>> -       { .name = "comm",       .entry =&sort_comm,     },
>> -       { .name = "dso",        .entry =&sort_dso,      },
>> -       { .name = "symbol",     .entry =&sort_sym,      },
>> -       { .name = "parent",     .entry =&sort_parent,   },
>> -       { .name = "cpu",        .entry =&sort_cpu,      },
>> +       { .name = "pid",        .entry =&sort_thread,                   },
>> +       { .name = "comm",       .entry =&sort_comm,                     },
>> +       { .name = "dso",        .entry =&sort_dso,                      },
>> +       { .name = "dso_from",   .entry =&sort_dso_from,.taken = true    },
>> +       { .name = "dso_to",     .entry =&sort_dso_to,   .taken = true   },
>> +       { .name = "symbol",     .entry =&sort_sym,                      },
>> +       { .name = "symbol_from",.entry =&sort_sym_from,.taken = true    },
>> +       { .name = "symbol_to",  .entry =&sort_sym_to,   .taken = true   },
>> +       { .name = "parent",     .entry =&sort_parent,                   },
>> +       { .name = "cpu",        .entry =&sort_cpu,                      },
>>
>> +       { .name = "mispredict", .entry =&sort_mispredict, },
>>  };
>
>
> The new sort dimensions don't seem to show up in perf report -h. Could you
> please update the help text?
>
Can do this with the understanding that those are only avail when you
use branch sampling.

> Also:
>
> # perf script -h
>
>    -f, --fields <str>    comma separated output fields prepend with 'type:'.
> Valid types: hw,sw,trace,raw. Fields:
> comm,tid,pid,time,cpu,event,trace,ip,sym,dso,addr
>
> You probably want to add a field here, so I could:
>
> perf record -b any_call,u -e cycles:u
> perf script -f event,branch_stack
>
> and examine raw (symbolized) samples like I can with
>
> perf record -g
> perf script -f event,ip,sym
>
Ok, I'll look into this. Looks like a useful command for automating processing.

>  -Arun
>
--
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/