Blame ChangeLogP510.txt

Packit 577717
2013-01-15  
Packit 577717
Packit 577717
  * 0917f567 src/threads.c: Cleaned up compiler warning (gcc version 4.4.6)
Packit 577717
Packit 577717
  * 06ca3faa src/components/bgpm/CNKunit/linux-CNKunit.c
Packit 577717
  src/components/bgpm/IOunit/linux-IOunit.c
Packit 577717
  src/components/bgpm/L2unit/linux-L2unit.c...: Cleaned up compiler warnings on
Packit 577717
  BG/Q (gcc version 4.4.6 (BGQ-V1R1M2-120920))
Packit 577717
Packit 577717
2013-01-14  
Packit 577717
Packit 577717
  * 56400627 .../build/lib.linux-x86_64-2.7/perfmon/__init__.py
Packit 577717
  .../lib.linux-x86_64-2.7/perfmon/perfmon_int.py
Packit 577717
  .../build/lib.linux-x86_64-2.7/perfmon/pmu.py...: libpfm4: remove extraneous
Packit 577717
  build artifacts.  Steve Kaufmann reported differences between the libpfm4 I
Packit 577717
  imported into PAPI and the libpfm4 that can be attained with a git clone
Packit 577717
  git://perfmon2.git.sourceforge.net/gitroot/perfmon2/libpfm4  Self: Do libpfm4
Packit 577717
  imports from a fresh clone of libpfm4.
Packit 577717
Packit 577717
2013-01-11  
Packit 577717
Packit 577717
  * 4ad994bc src/papi_events.csv: Clean up armv7 cortex a15 presets  Clean up
Packit 577717
  armv7 cortex a15 presets and add presets for L1 and L2 cache 
Packit 577717
Packit 577717
  * d54dabf5 ChangeLogP510.txt RELEASENOTES.txt doc/Doxyfile-common...: Prepare
Packit 577717
  the repo for a 5.1 release.  * Bump the version number to 5.1 * Update the
Packit 577717
  man pages * Create a changelog for 5.1 * Update RELEASENOTES
Packit 577717
Packit 577717
  * 8816a3b8 INSTALL.txt: Update INSTALL.txt  Add information about installing
Packit 577717
  PAPI on Intel MIC. Based upon information from Vince Weaver's PAPI MIC
Packit 577717
  support page. http://www.eece.maine.edu/~vweaver/projects/mic/
Packit 577717
Packit 577717
  * 8dc1ca23 TEST.TXT: Remove TEST.TXT  This was a leftover from a switch over
Packit 577717
  to git.
Packit 577717
Packit 577717
  * 292d6c9b src/papi_libpfm3_events.c: Fix build on ia64  When trying to build
Packit 577717
  papi 5.0.1 for IA64, my collegue got compile errors due to perfmon.h not
Packit 577717
  being included.  We're not sure if this actually is a configure bug, but this
Packit 577717
  patch fixed it.
Packit 577717
Packit 577717
  * 25424f41 src/extras.c: Fix kernel warning in _papi_hwi_stop_timer()  In
Packit 577717
  _papi_hwi_stop_timer() we were calling setitimer( timer, NULL, NULL ) to
Packit 577717
  disable the itimer.  Recent Linux kernels print warnings if you do this; NULL
Packit 577717
  is not a valid second argument to setitimer() and possibly this wasn't really
Packit 577717
  working before.  According to the manpage the proper fix is to call
Packit 577717
  setitimer() with a valid "new_value" field but with the values all 0.  That
Packit 577717
  is what this patch does.
Packit 577717
Packit 577717
2012-11-30  
Packit 577717
Packit 577717
  * a7d70127 src/components/micpower/README
Packit 577717
  src/components/micpower/Rules.micpower
Packit 577717
  src/components/micpower/linux-micpower.c...: MIC power component  The Intel
Packit 577717
  MIC (Xeon PHI) card reports power of several components of the card. These
Packit 577717
  values are reported in a sysfs file, so this component is cloned from the
Packit 577717
  coretemp component.
Packit 577717
Packit 577717
2013-01-08  
Packit 577717
Packit 577717
  * 121cd0a6 src/Makefile.in src/Rules.pfm4_pe src/configure...: configure: Add
Packit 577717
  shortcut for mic support.  * Add a --with-mic flag to enable the several
Packit 577717
  options to cross compile for mic. MIC builds are cross-compiled and Matt and
Packit 577717
  I were unable to figure out how to trigger cross compilation with just our
Packit 577717
  flag. This is short-hand for setting --with-arch=k1om --without-ffsll
Packit 577717
  --with-walltimer=clock_realtime_hr \ --with-perf-events --with-tls=__thread
Packit 577717
  --with-virtualtimer=cputime_id  * Automatically cause make to pass
Packit 577717
  CONFIG_PFMLIB_ARCH_X86=y to libpfm4's make.  So to build for the mic card one
Packit 577717
  has to do: {Set pathing to find the x86_64-k1om-linux-gcc cross-compiler} $
Packit 577717
  ./configure --host=x86_64-k1om-linux --with-mic $ make  Thanks to Matt
Packit 577717
  Johnson for the legwork on configure shortcuting.
Packit 577717
Packit 577717
2013-01-07  
Packit 577717
Packit 577717
  * f65c9d9e src/papi_events.csv: Add preset events for ARM Cortex A15
Packit 577717
Packit 577717
2012-12-14  
Packit 577717
Packit 577717
  * 61a9c7b1 man/man3/PAPI_get_eventset_component.3 src/papi.c: Doxygen: Add a
Packit 577717
  new API entry  Add the manpage for the new PAPI_get_eventset_component api
Packit 577717
  entry.
Packit 577717
Packit 577717
2013-01-02  
Packit 577717
Packit 577717
  * 38d969ab doc/Doxyfile-man1 doc/Doxyfile-man3 doc/Makefile...: Doxygen:
Packit 577717
  Cleanup generated man pages.  Mark a few \page sections as \htmlonly so that
Packit 577717
  man pages are not built for them. Modify the makefile to rm some data
Packit 577717
  structures that are generated.  Doxyfile-man3: * Take out papi_vector.h, this
Packit 577717
  file only defines a few data structures from which we don't need manpages. 
Packit 577717
  papi.h: * PAPI_get_component_index's inline comment had the close /**> to
Packit 577717
  delimit its description, but doxygen uses /**<.  papi_fwrappers.c: * Mark the
Packit 577717
  group PAPIF as internal so that a man page is not generated for it.  utils/*:
Packit 577717
  * Remove some useless htmlonly directrives, doxygen will generate pages for
Packit 577717
  any data structure, htmlonly doesn't stop that.  Doxyfile-man1: * Change a
Packit 577717
  flag in Doxyfile-man1 so that we don't document internal data structures in
Packit 577717
  the utilities.  We don't do this in -man3 because of the \class workaround we
Packit 577717
  use to create manpages for each of the PAPI_* api entry points. Because we
Packit 577717
  call them classes, they would be caught in the no data structures flag.
Packit 577717
Packit 577717
  * 7b790c09 doc/Doxyfile-html src/papi.h src/papi_fwrappers.c...: Doxygen:
Packit 577717
  Cleanup some of the markup  We were not using htmlonly correctly... The idea
Packit 577717
  was to use \htmlonly to not build manpages for a few things. To properly hide
Packit 577717
  \page s you want things like: /** \htmlonly \page Foo I don't want this to
Packit 577717
  generate a manpage. \endhtmlonly */
Packit 577717
Packit 577717
2012-12-07  
Packit 577717
Packit 577717
  * 152bac19 src/papi.c: Doxygen: Cleanup papi.c  Cleanup some \ref s, \ref
Packit 577717
  PAPI_function() isn't happy, use \ref PAPI_function it'll put in the proper
Packit 577717
  links.  Remove _papi_overflow_handler doc block. We had the block but no
Packit 577717
  code.
Packit 577717
Packit 577717
2012-12-20  
Packit 577717
Packit 577717
  * 7a40c769 src/components/rapl/tests/rapl_overflow.c: RAPL test code: Add
Packit 577717
  flexibility to the test code.  Per Will Cohen; ------------------ I was
Packit 577717
  reviewing some test results for the papi test and found that the
Packit 577717
  rapl_overflow.c tests makes an assumption that there are exactly two
Packit 577717
  packages. As a result the test will fail on machines with a single package. 
Packit 577717
  The following is a patch to make it a bit more flexible allow 1-n packages in
Packit 577717
  the test.  -Will -----------------
Packit 577717
Packit 577717
2012-12-19  
Packit 577717
Packit 577717
  * 96c9afb0 src/components/appio/README src/components/appio/appio.c
Packit 577717
  src/components/appio/appio.h...: Added events for seek statistics and support
Packit 577717
  for intercepting lseek() calls.
Packit 577717
Packit 577717
2012-12-14  
Packit 577717
Packit 577717
  * 003abf6d src/Rules.perfctr-pfm: Rules.perfctr-pfm: pass CC in all cases. 
Packit 577717
  Perfctr user library was not being passed CC when built.
Packit 577717
Packit 577717
2012-12-05  
Packit 577717
Packit 577717
  * e2c05b29 src/papi_internal.c: papi_internal.c: Refactor dublicated code in
Packit 577717
  cleanup and free eventset.  Currently the code to free runtime state is
Packit 577717
  duplicated in cleanup and free. The perf_event_uncore test exposed an issue
Packit 577717
  where free cleaned up cpu_attach state but cleanup did not, causing a leak. 
Packit 577717
  Have _papi_hwi_free_EventSet call _papi_hwi_cleanup_eventset to free most of
Packit 577717
  the runtime state of the eventset and then allow free_eventset to free the
Packit 577717
  Eventset Info struct.
Packit 577717
Packit 577717
2012-12-13  
Packit 577717
Packit 577717
  * 7d020224 src/configure src/configure.in: configure: Change fortran compiler
Packit 577717
  search order.  Bandaid fix to buildbot errors. By default, configure would
Packit 577717
  find icc before gcc but gfortran would be used before ifort.  The real fix is
Packit 577717
  to test that object code from the c compiler can be linked to by the fortran
Packit 577717
  compiler.
Packit 577717
Packit 577717
2012-12-12  
Packit 577717
Packit 577717
  * 87b6e913 src/papi_events.csv: ivy_bridge: remove PAPI_HW_INT event 
Packit 577717
  Apparently recent Intel Vol3B documentation removed this event, and the most
Packit 577717
  recent libpfm4 merge followed suit.  I asked at Intel about this and possibly
Packit 577717
  they only removed it because they didn't think anyone was using it.  Maybe
Packit 577717
  they'll ad it back
Packit 577717
Packit 577717
2012-12-10  
Packit 577717
Packit 577717
  * 293b26b9 src/Makefile.inc: Makefile.inc: Fix library link ordering.  Per
Packit 577717
  Will Cohen ----------------------------------------------------------- I ran
Packit 577717
  across a problem when trying to build papi with the bundled libpfm and an
Packit 577717
  earlier incompatible version of libpfm was already installed on the machine.
Packit 577717
  The make would use the /usr/lib{64}/libpfm.so before trying to use the
Packit 577717
  locally built version and this would cause problems.  The attached patch
Packit 577717
  changes the order of the linking and uses the local built libpfm before it
Packit 577717
  tries the installed version.  -Will
Packit 577717
  -----------------------------------------------------------
Packit 577717
Packit 577717
2012-12-12  
Packit 577717
Packit 577717
  * 57e6aa0d src/Makefile.in: Makefile.in: export CC_COMMON_NAME  In 17cfcb4a I
Packit 577717
  started using CC_COMMON_NAME in Rules.pfm4 but failed to have configure put
Packit 577717
  it in Makefile.
Packit 577717
Packit 577717
2012-12-11  
Packit 577717
Packit 577717
  * 17cfcb4a src/Rules.pfm4_pe src/configure src/configure.in: Cleanup icc
Packit 577717
  build  Start using -diag-disable to quiet down some of the remarks icc carps
Packit 577717
  about in libpfm4.  Also have configure export CC_COMMON_NAME and check
Packit 577717
  against that in Rules.pfm4_pe. afec8fc9a reverted us to passing
Packit 577717
  -Wno-unused-parameter to icc, polluting buildbot.
Packit 577717
Packit 577717
2012-12-10  
Packit 577717
Packit 577717
  * afec8fc9 src/configure src/configure.in: configure: Attempt to better
Packit 577717
  detect which C compiler we are using.  This attempts to address trac bug 162.
Packit 577717
  http://icl.cs.utk.edu/trac/papi/ticket/162  Specifying full paths for CC
Packit 577717
  caused issues in our configure logic. We set several flags specific to gcc or
Packit 577717
  icc and this was breaking down EG "/usr/bin/gcc" != "gcc"  Now we attempt to
Packit 577717
  execute whatever CC we are going to use and grep its version string. We set a
Packit 577717
  CC_COMMON_NAME \in {"gcc", "icc", "xlc", "unknown"} based upon the above and
Packit 577717
  later check CC_COMMON_NAME inplace of CC to set compiler specific flags.
Packit 577717
Packit 577717
  * 14432aa0 src/linux-timer.c src/papi.c: Minor Coverity fixes. Thanks, Will
Packit 577717
  Cohen.
Packit 577717
Packit 577717
2012-12-07  
Packit 577717
Packit 577717
  * ba5e83d4 src/papi_user_events.c: papi_user_events.c: Fix memory leak. 
Packit 577717
  Reported by William Cohen as detected by the coverity tool.
Packit 577717
Packit 577717
  * 166498a8 src/components/nvml/linux-nvml.c: nvml component: fix
Packit 577717
  detectDevices()  The routine detectDevices() always returned with the error
Packit 577717
  PAPI_ESYS when there was a device available. This resulted in that there were
Packit 577717
  no nvml events available. Fixed.
Packit 577717
Packit 577717
  * 11ad5894 src/components/nvml/linux-nvml.c: nvml component: add missing
Packit 577717
  variable declaration  In the routine _papi_nvml_init_componen(), the variable
Packit 577717
  papi_errorcode was not declared which prevented this component to build.
Packit 577717
  Added declaration of papi_errorcode as int.
Packit 577717
Packit 577717
2012-12-06  
Packit 577717
Packit 577717
  * 9567dfef src/ftests/first.F src/ftests/second.F: Fix warning messages
Packit 577717
  issued by gfortran 4.6.x regarding loss of precision when casting REAL to
Packit 577717
  INT. Thanks to Heike for identifying the proper intrinsics.
Packit 577717
Packit 577717
  * 72588227 src/papi.c src/papi.h: Add PAPI_get_eventset_component() to get
Packit 577717
  the component index from an event set. This is symmetric with
Packit 577717
  PAPI_get_event_component which extracts the information from an event. In
Packit 577717
  response to a request from John Mellor-Crummey.
Packit 577717
Packit 577717
  * 2e055d40 src/components/rapl/linux-rapl.c: Fix a compiler warning about a
Packit 577717
  possibly uninitialized return value.
Packit 577717
Packit 577717
2012-12-05  
Packit 577717
Packit 577717
  * 1aae2246 src/utils/command_line.c: Reformat the floating point output
Packit 577717
  string to recognize that you can't cast the *value* of a long long to a
Packit 577717
  double and expect to get the right answer; you need to cast the *pointer* to
Packit 577717
  a double, then everything works.
Packit 577717
Packit 577717
  * 0e834fc2 src/utils/command_line.c: Incorporated use of the new
Packit 577717
  PAPI_add_named_event API. Restructured output to support formatted printing
Packit 577717
  of built-in DATATYPEs: UINT64 prints as unsigned followed by (u); INT64
Packit 577717
  prints as signed; FP64 prints as float (but I don't like the default format);
Packit 577717
  BIT64 prints a hex, prefixed by '0x'. Also if info.units is not empty, units
Packit 577717
  are appended to output values. These features can be demo'd with the RAPL
Packit 577717
  component.
Packit 577717
Packit 577717
  * af6abec2 src/papi.h: Rearranged DATATYPE enums so INT64 is now default (0)
Packit 577717
  value. Also added a BIT64 type for unspecified bitfields.
Packit 577717
Packit 577717
2012-12-04  
Packit 577717
Packit 577717
  * 862033e0 src/components/bgpm/IOunit/linux-IOunit.c
Packit 577717
  src/components/bgpm/IOunit/linux-IOunit.h
Packit 577717
  src/components/bgpm/L2unit/linux-L2unit.c...: Resolved multiple components
Packit 577717
  conflict on BG/Q when overflow is enabled for multiple events from different
Packit 577717
  components at the same time.
Packit 577717
Packit 577717
  * 44744002 src/utils/command_line.c: Add -x and -u options to
Packit 577717
  papi_command_line to allow printing counter values in hexadecimal and
Packit 577717
  unsigned formats.
Packit 577717
Packit 577717
2012-11-30  
Packit 577717
Packit 577717
  * 25a914c5 src/papi_user_events.c: Cleanup unused variable warnings in
Packit 577717
  user_events code.
Packit 577717
Packit 577717
2012-11-28  
Packit 577717
Packit 577717
  * 9a75f872 src/Rules.pfm4_pe src/configure src/configure.in: Cleanup the
Packit 577717
  build under icc.  libpfm4's build system uses a gcc specific flag,
Packit 577717
  -Wno-unused-parameter. It does this via a variable, DBG, in config.mk:
Packit 577717
  DBG?=-g -Wall -Werror -Wextra -Wno-unused-parameter  The Intel compiler
Packit 577717
  doesn't understand -Wno-unused-parameter and complains about it. In
Packit 577717
  Rules.pfm4_pe we set DBG for icc builds.
Packit 577717
Packit 577717
2012-11-27  
Packit 577717
Packit 577717
  * 4def827b src/configure src/configure.in: Fix the perfctr build that was
Packit 577717
  breaking due to missing CPU  Mark Gates was reporting PAPI 5 wasn't running
Packit 577717
  properly on Keeneland.  It looks like some CPU cleanups in the configure code
Packit 577717
  broke things.  Hopefully this helps the situation.
Packit 577717
Packit 577717
2012-11-21  
Packit 577717
Packit 577717
  * 4316f172 src/perf_events.c: perf_events: get rid of "PAPI Error: Didn't
Packit 577717
  close all events" error  This was more meant as a warning; it could trigger
Packit 577717
  when closing an EventSet that had an event partially added but failed for
Packit 577717
  some reason.
Packit 577717
Packit 577717
  * 671e10bd src/utils/command_line.c: papi_command_line: fix error output  The
Packit 577717
  error messages got a bit weird looking due to the PAPI error printing changes
Packit 577717
  a while back.
Packit 577717
Packit 577717
  * 959afa49 src/papi_internal.c: Fix _papi_hwi_add_event to report errors back
Packit 577717
  to user.  Previously _papi_hwi_add_event would report all errors returned by
Packit 577717
  add_native_events() as being PAPI_ECNFLCT even though add_native_events()
Packit 577717
  returned a wider range of errors.
Packit 577717
Packit 577717
  * 8ecb70ba src/perf_events.c: Have perf_event return PAPI_EPERM rather than
Packit 577717
  PAPI_ECNFLCT if the kernel itself returns EPERM
Packit 577717
Packit 577717
  * 9053ca1c src/perf_events.c: Work around kernel issue with
Packit 577717
  PERF_EVENT_IOC_REFRESH  It's unclear exactly the best way to restart
Packit 577717
  sampling.  Refreshing with 1 is the "official" way as espoused by the kernel
Packit 577717
  developers, but it doesn't work on Power.  0 works for Power and most other
Packit 577717
  machines, but the kernel developers say not to use it.  This makes power user
Packit 577717
  0 until we can figure out exactly what is going on.
Packit 577717
Packit 577717
  * e85df04b src/components/appio/tests/appio_test_socket.c: - added support
Packit 577717
  distinguishing between network and file I/O. - added events to measure
Packit 577717
  statistics for sockets - updated README
Packit 577717
Packit 577717
2012-11-15  
Packit 577717
Packit 577717
  * 248694ef src/x86_cpuid_info.c: Update x86_cpuid_info code for KNC.  On
Packit 577717
  Knight's Corner the leaf2 code returns 0 for the count value. We were
Packit 577717
  printing a warning on this; better would be to just skip the cache detection
Packit 577717
  code if we get this result.
Packit 577717
Packit 577717
2012-11-08  
Packit 577717
Packit 577717
  * 82c93156 src/linux-bgp-memory.c src/linux-bgp.c src/linux-bgp.h: There was
Packit 577717
  more cleaning up necessary in order to get PAPI compiled on BG/P. It should
Packit 577717
  work now with the recommended configure steps described in INSTALL.
Packit 577717
Packit 577717
2012-11-07  
Packit 577717
Packit 577717
  * 77da80b3 src/Makefile.inc src/configure src/configure.in...: Make BGP use
Packit 577717
  papi_events.csv  This was easier than trying to clean up the
Packit 577717
  linux-bgp-preset-events.c file to have the proper file layout.
Packit 577717
Packit 577717
  * fc8a4168 src/linux-bgp.c: Fix some linux-bgp build issues.  No one has
Packit 577717
  tried compiling after all the PAPI 5.0 changes so many bugs slipped in.
Packit 577717
Packit 577717
  * c16ef312 src/ctests/perf_event_uncore.c: Fix type warnings in
Packit 577717
  perf_event_uncore test.
Packit 577717
Packit 577717
  * 3947e9c8 src/ctests/perf_event_uncore.c: Put a bandaid on the
Packit 577717
  perf_event_uncore test.  Check for an Intel family 6 model 45 processor
Packit 577717
  (sandybridge ep) before executing the test.
Packit 577717
Packit 577717
2012-09-27  
Packit 577717
Packit 577717
  * a23d95f8 src/papi.c src/papi.h src/papi_fwrappers.c...: Mark some comments
Packit 577717
  @htmlonly.  This cleans up what man pages are generated.
Packit 577717
Packit 577717
2012-11-07  
Packit 577717
Packit 577717
  * d239c350 src/Makefile.inc src/Rules.pfm4_pe: Factor out duplicate install
Packit 577717
  code from Rules.pfm4_pe  The Makefile.inc has a rule to installed shared
Packit 577717
  libraries.  However, Rules.pfm4_pe also has a slightly different set of rules
Packit 577717
  to install code for shared libraries.  This leads to the same shared library
Packit 577717
  being installed under two different names.  The duplicate code has been
Packit 577717
  removed from Rules.pfm4_pe and a symbolic link has been added to ensure that
Packit 577717
  any code that might have linked with
Packit 577717
  libpapi.so.$(PAPIVER).$(PAPIREV).$(PAPIAGE) still runs. 
Packit 577717
Packit 577717
2012-10-30  
Packit 577717
Packit 577717
  * fcc64ff9 src/papi_events.csv: Add PAPI_HW_INT event for IvyBridge
Packit 577717
Packit 577717
2012-10-26  
Packit 577717
Packit 577717
  * ef89fc56 src/papi_events.csv: MIC: update PAPI_FP_INS / PAPI_VEC_INS
Packit 577717
  instruction  We were using VPU_INSTRUCTIONS_EXECUTED for PAPI_FP_INS but
Packit 577717
  really it's more appropriate for PAPI_VEC_INS  This leaves PAPI_FP_INS
Packit 577717
  undefined, which breaks a lot of the ctests. A long term goal should probably
Packit 577717
  be modifying the tests to use another counter if PAPI_FP_INS isn't available
Packit 577717
  (this affects Ivy Bridge too).
Packit 577717
Packit 577717
2012-10-25  
Packit 577717
Packit 577717
  * 975c03f1 src/perf_events.c: perf_event: fix granularity bug  cut-and paste
Packit 577717
  error in the last set of changes.  Would have meant if you tried to
Packit 577717
  explicitly set granularity to thread you'd get system instead.
Packit 577717
Packit 577717
  * 3cd3a62d src/configure src/configure.in src/ctests/Makefile...: Add
Packit 577717
  perf_event_uncore ctest  Also add a new type of ctest, perf_event specific 
Packit 577717
  In theory we should have configure only enable this if perf_event support is
Packit 577717
  being used.
Packit 577717
Packit 577717
  * 5ee97430 src/perf_events.c: perf_event: add PAPI_DOM_SUPERVISOR to allowed
Packit 577717
  perf_event domains  perf_event supports this domain but since we didn't have
Packit 577717
  it in the list PAPI wasn't letting us set/unset this.  This is needed for
Packit 577717
  uncore support, as for uncore domain must be set to allow monitoring
Packit 577717
  everything.
Packit 577717
Packit 577717
  * c9325560 src/perf_events.c: perf_event enable granularity support  Add
Packit 577717
  support for PAPI_GRAN_SYS to perf_event. This is needed for uncore support.
Packit 577717
Packit 577717
2012-10-18  
Packit 577717
Packit 577717
  * 59d3d758 src/mb.h src/perf_events.c: Update the memory barriers  It turns
Packit 577717
  out PAPI fails on older 32-bit x86 machines because it tries to use an SSE
Packit 577717
  rmb() memory barrier.  (Yes, I'm trying to run PAPI on a Pentium II.  Don't
Packit 577717
  ask)  It looks like our memory barriers were copied out of the kernel, which
Packit 577717
  doesn't quite work because it expects some kernel infrastructure instead. 
Packit 577717
  This patch uses the definitions used by the "perf" tool instead.  Also
Packit 577717
  dropped the use of the mb() memory barrier on mmap tail write, as the perf
Packit 577717
  tool itself did a while ago so I'm hoping it's safe to do so as well.  It
Packit 577717
  makes these definitions a lot simpler.
Packit 577717
Packit 577717
2012-10-08  
Packit 577717
Packit 577717
  * bcdce5bc src/perf_events.c: perf_event: clarify an error message  The
Packit 577717
  message was saying detecting rdpmc support broke, but the real error is that
Packit 577717
  perf_events itself is totally broken on this machine and it's just rdpcm was
Packit 577717
  the first code that tried to access it.
Packit 577717
Packit 577717
2012-10-02  
Packit 577717
Packit 577717
  * 3bb3558f src/mb.h: Update memory barries for Knights Corner  Despite being
Packit 577717
  x86_64 they don't support the SSE memory barrier instructions, so add a case
Packit 577717
  in mb.h to handle this properly.
Packit 577717
Packit 577717
2012-10-01  
Packit 577717
Packit 577717
  * 38a5d74c src/libpfm4/README src/libpfm4/docs/Makefile
Packit 577717
  src/libpfm4/docs/man3/libpfm_intel_atom.3...: Merge libpfm4 with Knights
Packit 577717
  Corner Support
Packit 577717
Packit 577717
  * bf959960 src/papi_events.csv: Change "phi" to "knc" to match libpfm4 for
Packit 577717
  Xeon Phi / Knights Corner support
Packit 577717
Packit 577717
2012-09-20  
Packit 577717
Packit 577717
  * d9249635 ChangeLogP501.txt RELEASENOTES.txt: Update releasenotes and add a
Packit 577717
  changelog for 5.0.1
Packit 577717
Packit 577717
  * a1e30348 man/man1/papi_avail.1 man/man1/papi_clockres.1
Packit 577717
  man/man1/papi_command_line.1...: Rebuild the manpages for a 5.0.1 release.