> Date: Tue, 5 Jul 2011 08:18:48 -0500
> From: maynardj@us.ibm.com
> To: duqisi@hotmail.com
> CC: oprofile-list@lists.sourceforge.net
> Subject: Re: opreport error with "-m cpu"
>
> On 07/04/2011 10:16 AM, DuEugene wrote:
> > Hi all,
> > I am using oprofile on an SMP system,with a setting of "--separate=all".When
> > sampling finished, I use "opreport -l -m cpu" to analyse the data ,and I got a
> > error message "opreport error: sanitize_profile_sample_files(): sample \ file
> > parsed twice".When I checked the source code of oprofile, I found that the
> > method named "parsed_filename::profile_spec_equal(parsed_filename const &
> > parsed)" in parse_filename.cpp seemed to be a little strange.
> > bool parsed_filename::profile_spec_equal(parsed_filename const & parsed)
> > {
> > return event == parsed.event &&
> > count == parsed.count &&
> > unitmask == parsed.unitmask &&
> > tgid == parsed.tgid &&
> > tid == parsed.tid &&
> > cpu == parsed.tid;
> > }
> >
> > I wonder if "cpu == parsed.tid" here should be "cpu == parsed.cpu"?
> Seems likely. Can you patch the source and run the patched opreport to see if
> that resolves it?
>
> Thanks.
> -Maynard

Yes,it works.

diff -Naur oprofile-0.9.7-rc2/libpp/parse_filename.cpp oprofile-0.9.7-rc2_new/libpp/parse_filename.cpp
--- oprofile-0.9.7-rc2/libpp/parse_filename.cpp 2011-07-05 10:25:04.000000000 +0800
+++ oprofile-0.9.7-rc2_new/libpp/parse_filename.cpp     2011-07-15 16:52:48.000000000 +0800
@@ -237,7 +237,7 @@
                unitmask == parsed.unitmask &&
                tgid == parsed.tgid &&
                tid == parsed.tid &&
-               cpu == parsed.tid;
+               cpu == parsed.cpu;
 }
 
 ostream & operator<<(ostream & out, parsed_filename const & data)


>
> > oprofile version is 0.9.5, the source code of version 0.9.7 is also the seem.
> >
> >
> >
> >
> > ------------------------------------------------------------------------------
> > All of the data generated in your IT infrastructure is seriously valuable.
> > Why? It contains a definitive record of application performance, security
> > threats, fraudulent activity, and more. Splunk takes this data and makes
> > sense of it. IT sense. And common sense.
> > http://p.sf.net/sfu/splunk-d2d-c2
> >
> >
> >
> > _______________________________________________
> > oprofile-list mailing list
> > oprofile-list@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/oprofile-list
>