On 19.06.12 08:09:31, Stephane Eranian wrote:

On Tue, Jun 19, 2012 at 4:13 AM, Namhyung Kim [off-list ref] wrote:

quoted

Hi, Robert

On Mon, 18 Jun 2012 16:51:08 +0200, Robert Richter wrote:

quoted

On 29.09.11 18:48:01, Arnaldo Carvalho de Melo wrote:

quoted

From: Jim Cromie <jim.cromie@gmail.com>

This perf stat option emulates valgrind's --log-fd option, allowing the
user to send perf results elsewhere, and leaving stderr for use by the
program under test. �This complements --output file option, and is
mutually exclusive with it.

� �3>results �perf stat --log-fd 3 � � � � �-- $cmd
� �3>>results perf stat --log-fd 3 --append -- $cmd

The perl distro's make test.valgrind target uses valgrind's --log-fd
option, I've adapted it to invoke perf also, and tested this patch
there.

Link: http://lkml.kernel.org/r/1315437244-3788-2-git-send-email-jim.cromie@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <redacted>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <redacted>

With certain shell redirections this (56f3bae) fails with a log fd
setup failure. Fix below.

It looks somewhat related to Stephane's patch. Can you we check it too?

http://lkml.indiana.edu/hypermail/linux/kernel/1205.1/04232.html

I applied the patch to my tree and this works too. Both patches
basically avoid fdopen() if a log-fd option is not given.

After reviewing the code again I noticed my code breaks the -o option.
So Stephane's patch is fine for me too, except for the fact that
--log-fd 0 is ignored. Maybe we change this by initializing output_fd
with -1 and modify the checks of output_fd?

Yes, my patch is needed and it was posted over a month ago now.....
Arnaldo, please apply my patch.

It is in the pull request for Ingo.

-Robert

-- 
Advanced Micro Devices, Inc.
Operating System Research Center