linux-kernel - Re: [PATCH] perf trace: Simplify '--summary' output

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, 12 Nov 2013 22:33:34 +0100
From:	Ingo Molnar <mingo@...nel.org>
To:	Pekka Enberg <penberg@...nel.org>
Cc:	linux-kernel@...r.kernel.org, David Ahern <dsahern@...il.com>,
	Ingo Molnar <mingo@...hat.com>,
	Arnaldo Carvalho de Melo <acme@...stprotocols.net>
Subject: Re: [PATCH] perf trace: Simplify '--summary' output


* Pekka Enberg <penberg@...nel.org> wrote:

> The output of 'perf trace --summary' tries to be too cute with
> formatting and makes it very hard to read.  Simplify it in the spirit of
> "strace -c":
> 
> [penberg@...alhost libtrading]$ perf trace -a --duration 10000 --summary -- sleep 1
> ^C

Btw., just a very small feature suggestion - instead of forcing the 
somewhat counter-intuitive '-- sleep X' pattern it might be helpful to 
have a general option parser for 'delay' values, which would allow a wide 
range of time units like:

	--delay 10.5h
        --delay 0.1s
        --delay 0.1sec

etc. - with the default unit being 'seconds'.

That parser could then be applied to '--duration' filter as well, allowing 
things like:

   --duration 10sec

which IMHO are more obvious to read (and easier to remember!) than 
'--duration 10000'.

>  Summary of events:
> 
>  dbus-daemon (555), 10 events, 0.0%, 0.000 msec
> 
>                                                     msec/call
>    syscall            calls      min      avg      max stddev
>    --------------- -------- -------- -------- -------- ------
>    sendmsg                2    0.002    0.005    0.008  55.00
>    recvmsg                2    0.002    0.003    0.005  44.00
>    epoll_wait             1    0.000    0.000    0.000   0.00

In what units is stddev? Percentage? If yes then it might be useful to 
output it as %. [and probably to restrict precision to a single digit, 
standard deviations are rarely more accurate than 0.1%.]

>  NetworkManager (667), 56 events, 0.0%, 0.000 msec
> 
>                                                     msec/call
>    syscall            calls      min      avg      max stddev
>    --------------- -------- -------- -------- -------- ------
>    poll                   2    0.000    0.002    0.003 100.00
>    sendmsg               10    0.004    0.007    0.016  15.41
>    recvmsg               16    0.002    0.003    0.005   8.24
> 
>  zfs-fuse (669), 4 events, 0.0%, 0.000 msec
> 
>                                                     msec/call
>    syscall            calls      min      avg      max stddev
>    --------------- -------- -------- -------- -------- ------
>    futex                  2    0.000    0.001    0.002 100.00

Nice looking output btw! :-)

Acked-by: Ingo Molnar <mingo@...nel.org>

Thanks,

	Ingo
--
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/