0.6.1 release candidate

[John L. <le...@mo...>]

http://movementarian.org/oprofile-0.6.1.tar.gz

Please test it out. Release notes below.

regards
john


OProfile is a powerful system-wide profiler for Linux. Read
more at http://oprofile.sf.net

OProfile 0.6.1 has been released. OProfile is still in alpha,
but has been proven stable for many users.

Release notes
-------------

For 2.2 kernels, the module must be compiled as the same user
that owns the kernel source tree.
 
nosmp is not supported in kernels before 2.4.10 (bug #463087).

The pre-emptable kernel option is not supported in 2.4 (bug #478516).
 
Power management on laptops can be incompatible with OProfile in 2.4 (bug #554927).

New features
------------

The opannotate option --source-dir has been dropped. In its place are 
--search-dirs and --base-dirs, which allows more flexibility in locating
source files for annotation. Annotated output files are always generated
with the full path.

A feature present in the old oprofpp has been re-added for opreport: you can
now compare samples counts side-by-side for more than one event via
profile specifications such as "event:DATA_MEM_REFS,CPU_CLK_UNHALTED".

Bug fixes
---------

The opcontrol(1) man page has been updated for the --event changes in
the last release.

opcontrol now correctly defaults to the default unit mask if one is
not specified by the user. This was most commonly seen as Pentium IV
users receiving no samples.

A workaround for a GCC debug info problem (bug 11744) has been added to
prevent oprofile thinking functions that have debug info do not.

Slow behaviour of opcontrol --dump has been improved.

Some minor smart demangling improvements have been made.

The daemon is now more robust against opcontrol --reset races.

The 2.6 system call has been fixed for s390.

Failure to open binary files is no longer fatal in opreport, generating
a warning instead.

Fix mis-credited samples when two samples going to the same shared library are
separated by a context switch.

Prevent the user from using exactly the same event for two counters as
it caused a daemon segfaullt.
-- 
Khendon's Law:
If the same point is made twice by the same person, the thread is over.