Bug 748789

Description of problem:
Some of our tests for oprofile regularly fail on certain HW, so I start reporting these so either our tests can be fixed (if they are wrong), or the genuine bug in oprofile is confirmed. 

Some of our tests work try to collect some data in oprofile and then run some commands working with data dumped. This usually works, but sometimes this fails, and it is usually bound to a certain HW or certain box.

The test which fails follows a oprofile tutorial for Fedora 12, on http://people.redhat.com/wcohen/Fedora12OprofileTutorial.txt

::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
:: [   LOG    ] :: Oprofile configuration
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::

:: [   PASS   ] :: Running 'opcontrol --init'
:: [07:21:48] ::  x86-64/family10
:: [   PASS   ] :: Running 'opcontrol --setup --no-vmlinux'
:: [   PASS   ] :: Running 'ophelp --get-default-event'
:: [07:21:48] ::  Using CPU_CLK_UNHALTED:100000:0:1:1 as default event
:: [   PASS   ] :: Running 'opcontrol --setup --no-vmlinux --event=CPU_CLK_UNHALTED:100000:0:1:1'
:: [   PASS   ] :: Running 'opcontrol --setup --vmlinux=/usr/lib/debug/lib/modules/2.6.32-211.el6.i686/vmlinux --event=CPU_CLK_UNHALTED:100000:0:1:1'
Using 2.6+ OProfile kernel interface.
Reading module info.
Using log file /var/lib/oprofile/samples/oprofiled.log
Daemon started.
Profiler running.
:: [   PASS   ] :: Running 'opcontrol --start'
:: [   PASS   ] :: Running 'ps aux | grep oprofiled'
:: [   PASS   ] :: Running 'opcontrol --shutdown'

::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
:: [   LOG    ] :: Collecting and analysing data
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::

:: [   PASS   ] :: Running 'opcontrol --start'
:: [07:21:56] ::  Sleeping for a while
:: [   PASS   ] :: Running 'opcontrol --dump'
:: [   PASS   ] :: File /var/lib/oprofile/samples/current should exist
:: [   PASS   ] :: Running 'opcontrol --shutdown'
:: [   PASS   ] :: Running 'opcontrol --save=junk'
:: [   PASS   ] :: Running 'opcontrol --start'
:: [07:22:08] ::  Executing the benchmark
<output snipped>
:: [07:22:10] ::  Benchmark finished
:: [   PASS   ] :: Running 'opcontrol --dump'
:: [07:22:11] ::  Sleeping for a while
opreport error: No sample file found: try running opcontrol --dump
or specify a session containing sample files

:: [   FAIL   ] :: Running 'opreport --long-filenames --threshold 1' (Expected 0, got 1)
<same error>
:: [   FAIL   ] :: Running 'opreport --long-filenames --threshold 1 | grep linpack1000d' (Expected 0, got 1)
<same error>
:: [   FAIL   ] :: Running 'opreport -l image:/mnt/testarea/tests/oprofile/Sanity/fedora12-tutorial/linpack1000d --threshold .1' (Expected 0, got 1)
<same error>
:: [   FAIL   ] :: Running 'opreport -l image:/mnt/testarea/tests/oprofile/Sanity/fedora12-tutorial/linpack1000d --threshold .1 | grep daxpy' (Expected 0, got 1)
<same error>
:: [   FAIL   ] :: Running 'opreport -l image:/mnt/testarea/tests/oprofile/Sanity/fedora12-tutorial/linpack1000d --include-symbols=daxpy_ --details' (Expected 0, got 1)
<same error>
:: [   FAIL   ] :: Running 'opannotate --source /mnt/testarea/tests/oprofile/Sanity/fedora12-tutorial/linpack1000d -o session_annote' (Expected 0, got 1)
:: [   FAIL   ] :: File session_annote should exist 
:: [   PASS   ] :: Running 'opcontrol --shutdown'

::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
:: [   LOG    ] :: Saving profiling data
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::

:: [   PASS   ] :: Running 'opcontrol --start'
:: [07:22:28] ::  Executing the benchmark
<output snipped>
:: [07:22:30] ::  Benchmark finished
:: [   PASS   ] :: Running 'opcontrol --save=linpack'
:: [   PASS   ] :: Running 'opcontrol --dump'
:: [07:22:33] ::  Sleeping for a while
error: no sample files found: profile specification too strict ?
:: [   FAIL   ] :: Running 'opreport session:linpack --long-filenames --threshold 1' (Expected 0, got 1)
error: no sample files found: profile specification too strict ?
:: [   FAIL   ] :: Running 'opreport session:linpack --long-filenames --threshold 1 | grep linpack1000d' (Expected 0, got 1)
error: no sample files found: profile specification too strict ?
:: [   FAIL   ] :: Running 'opreport session:linpack -l image:/mnt/testarea/tests/oprofile/Sanity/fedora12-tutorial/linpack1000d --threshold .1' (Expected 0, got 1)
error: no sample files found: profile specification too strict ?
:: [   FAIL   ] :: Running 'opreport session:linpack -l image:/mnt/testarea/tests/oprofile/Sanity/fedora12-tutorial/linpack1000d --threshold .1 | grep daxpy' (Expected 0, got 1)
error: no sample files found: profile specification too strict ?
:: [   FAIL   ] :: Running 'opreport session:linpack -l image:/mnt/testarea/tests/oprofile/Sanity/fedora12-tutorial/linpack1000d --include-symbols=daxpy_ --details' (Expected 0, got 1)
error: no sample files found: profile specification too strict ?
:: [   FAIL   ] :: Running 'opannotate session:linpack --source /mnt/testarea/tests/oprofile/Sanity/fedora12-tutorial/linpack1000d -o linpack_annote' (Expected 0, got 1)
:: [   FAIL   ] :: File linpack_annote should exist

Version-Release number of selected component (if applicable):
oprofile-0.9.6-21.el6.i686

How reproducible:
always on hp-bl495cg5-01.rhts.bos.redhat.com (4cores, family 16=10h, model 2, arch i686)

Steps to Reproduce:
1. see above
  
Actual results:
opreport and opannotate do not show anything after opcontrol --dump

Expected results:
something useful

Additional info:
This is very environment specific. The identical test PASSed od different box with same CPU family/model running x86_64. I was not able to test if the real factor is the box itself, or x86-64ness.

This is not a regression.