http://blog.gmane.org/gmane.linux.oprofile hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://comments.gmane.org/gmane.linux.oprofile/11531 <pre>Add support for IBM POWER8 processor The Power ISA 2.07 was recently published at http://power.org/documentation. The IBM POWER8 processor currently under development is an implementation of the ISA 2.07. This patch provides the initial support for POWER8 to oprofile. NOTE: Only operf support is included with this patch. The legacy opcontrol-based profiler and the oprofile kernel module have not been (and may never be) updated to support this new processor. Signed-off-by: Maynard Johnson &lt;maynardj&lt; at &gt;us.ibm.com&gt; --- events/Makefile.am | 1 + events/ppc64/power8/events | 97 ++++++++++++++++++++++++++++++++++++++++ events/ppc64/power8/unit_masks | 9 ++++ libop/op_cpu_type.c | 1 + libop/op_cpu_type.h | 1 + libop/op_events.c | 1 + utils/opcontrol | 5 ++ utils/ophelp.c | 4 +- 8 files changed, 118 insertions(+), 1 deletions(-) create mode 100644 events/ppc64/power8/events create mode 100644 events/pp</pre> Maynard Johnson 2013-05-24T15:35:20 http://comments.gmane.org/gmane.linux.oprofile/11529 <pre>Hello, I'm interested in using OProfile for PowerPC processors, in particular the MPC853T running on Linux Kernel 2.4. Is OProfile compatible with this configuration? The MPC853T belongs to the MPC8xx series that use the PowerQuicc I architecture. The documentation says that there's limited support for the 32 bit PowerPCs in particular for the 2.4 kernel, so I'd like to know if this includes that processor and if possible I'd also like to know a bit more about those limitations. Can someone help me? Thanks in advance João Cabeleira ------------------------------------------------------------------------------ Try New Relic Now &amp; We'll Send You this Cool Shirt New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, &amp; servers with just a few lines of code. Try New Relic and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may_______________________________________________ oprofile-</pre> Cabeleira, Joao (EXT-RandstadTechnologies - PT/Amadora 2013-05-24T12:18:20 http://comments.gmane.org/gmane.linux.oprofile/11526 <pre>This patch fixes a bug if daemonrc file doesn't exist or doesn't contain a 'BUF_SIZE=' line : initalizes 'buffer_size' to the default value to prevent unpredictable value Signed-off-by: Gilles Allard &lt;gilles.allard21&lt; at &gt;orange.fr&gt; --- gui/oprof_start_config.cpp | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/gui/oprof_start_config.cpp b/gui/oprof_start_config.cpp index 02bfe64..40c6e3d 100644 --- a/gui/oprof_start_config.cpp +++ b/gui/oprof_start_config.cpp &lt; at &gt;&lt; at &gt; -34,6 +34,7 &lt; at &gt;&lt; at &gt; event_setting::event_setting() config_setting::config_setting() : +buffer_size(OP_DEFAULT_BUF_SIZE), no_kernel(false), verbose(false), separate_lib(false), </pre> Gilles Allard 2013-05-23T15:19:39 http://comments.gmane.org/gmane.linux.oprofile/11525 <pre>This patch fixes error messages send by Qt4 ( Qt4-4.8.2 ) during oprof_start startup. This is a minor bug: errors detected by Qt don't seem to prevent oprof_start from working Signed-off-by: Gilles Allard &lt;gilles.allard21&lt; at &gt;orange.fr&gt; --- gui/oprof_start.cpp | 12 ++++++------ gui/oprof_start.h | 10 +++++----- gui/ui/oprof_start.base.ui | 28 ++++++++++++++-------------- 3 files changed, 25 insertions(+), 25 deletions(-) diff --git a/gui/oprof_start.cpp b/gui/oprof_start.cpp index 7f46a18..daa513d 100644 --- a/gui/oprof_start.cpp +++ b/gui/oprof_start.cpp &lt; at &gt;&lt; at &gt; -843,7 +843,7 &lt; at &gt;&lt; at &gt; uint oprof_start::max_perf_count() const } -void oprof_start::on_flush_profiler_data() +void oprof_start::on_flush_profiler_data_btn_clicked(bool /* checked */) { vector&lt;string&gt; args; args.push_back("--dump"); &lt; at &gt;&lt; at &gt; -856,7 +856,7 &lt; at &gt;&lt; at &gt; void oprof_start::on_flush_profiler_data() // user is happy of its setting. -void oprof_start::on_start_profiler() +void oprof_start::on_start_profiler_btn_clicked(bool /</pre> Gilles Allard 2013-05-23T15:14:14 http://comments.gmane.org/gmane.linux.oprofile/11524 <pre>Hello, This is my first post here. Sorry if the issue is to basic for this list. I am tiring to run oprofile on a Blackfin uClinux based system. I have compiled the oprofile as it is described in the link http://docs.blackfin.uclinux.org/doku.php?id=linux-kernel:oprofile&amp;s[]=oprofile I am compiling some external (from uClinux) software components like the telephony components Asterisk DAHDI etc. When I try to run oprofile I get root&lt; at &gt;pr1:~&gt; /usr/bin/bfin_opcontrol --init mkdir: can't create directory '/var/lib/': File exists root&lt; at &gt;pr1:~&gt; /usr/bin/bfin_opcontrol --start-daemon &amp; [1] 469 /usr/bin/bfin_opcontrol --start-daemon root&lt; at &gt;pr1:~&gt; Using 2.6+ OProfile kernel interface. Reading module info. Using log file /var/lib/oprofile/oprofiled.log [1] Done /usr/bin/bfin_opcontrol --start-daemon root&lt; at &gt;pr1:~&gt; /usr/bin/bfin_opcontrol --start root&lt; at &gt;pr1:~&gt; /usr/bin/bfin_opcontrol --dump root&lt; at &gt;pr1:~&gt; /usr/bin/opreport -l /usr/bin/opreport error: /var/lib/oprofile/samples/current/{kern}/dahdi/{dep}/{ke</pre> Dimitar Penev 2013-05-23T07:00:24 http://comments.gmane.org/gmane.linux.oprofile/11521 <pre>Fix recent regression involving unit mask values of '0' A patch that I wrote and committed on Jan 23, 2013 broke unit mask handling for mask values of '0'. As far as I can tell, i386/core_2 is the only cpu type that has unit masks with values of '0'. So, for example, the following command will fail as shown below: operf -e L2_M_LINES_OUT:100000:0 ls Invalid unit mask 0x0 for event L2_M_LINES_OUT Error retrieving info for event L2_M_LINES_OUT:100000:0 This patch fixes that problem. Signed-off-by: Maynard Johnson &lt;maynardj&lt; at &gt;us.ibm.com&gt; --- libop/op_events.c | 8 ++++++++ 1 files changed, 8 insertions(+), 0 deletions(-) diff --git a/libop/op_events.c b/libop/op_events.c index e5ecbcc..b82128e 100644 --- a/libop/op_events.c +++ b/libop/op_events.c &lt; at &gt;&lt; at &gt; -1064,6 +1064,14 &lt; at &gt;&lt; at &gt; static int _is_um_valid_bitmask(struct op_event * event, u32 passed_um) } evt.unit = tmp_um_no_dups; +// Now check if passed um==0 and if the defined event has a UM with value '0'. +if (!passed_um) { +for (i = 0; i &lt; </pre> Maynard Johnson 2013-05-22T14:11:45 http://comments.gmane.org/gmane.linux.oprofile/11518 <pre>I'd like to revive this old patch from 2008: diff -u a/utils/opcontrol b/utils/opcontrol --- a/utils/opcontrol 2008-11-11 14:31:25.000000000 +0100 +++ b/utils/opcontrol 2008-11-11 12:30:42.000000000 +0100 &lt; at &gt;&lt; at &gt; -187,7 +187,7 &lt; at &gt;&lt; at &gt; fi fi mkdir /dev/oprofile &gt;/dev/null 2&gt;&amp;1 - grep oprofilefs /etc/mtab &gt;/dev/null + grep oprofilefs /proc/mounts &gt;/dev/null if test "$?" -ne 0; then mount -t oprofilefs nodev /dev/oprofile &gt;/dev/null fi &lt; at &gt;&lt; at &gt; -1607,7 +1607,7 &lt; at &gt;&lt; at &gt; do_deinit() { # unmount /dev/oprofile if it is mounted - OPROF_FS=`grep /dev/oprofile /etc/mtab` + OPROF_FS=`grep /dev/oprofile /proc/mounts` if test -n "$OPROF_FS"; then umount /dev/oprofile fi &lt; at &gt;&lt; at &gt; -1705,7 +1705,7 &lt; at &gt;&lt; at &gt; check_version() { OPROFILE_AVAILABLE=no - grep oprofilefs /etc/mtab &gt;/dev/null </pre> Mark Pearson 2013-05-21T19:54:33 http://comments.gmane.org/gmane.linux.oprofile/11513 <pre>Fix Coverity errors found on May 20, 2013 git snapshot Coverity identified the following errors on scans run from May 7 through May 20, 2013: Type,Category,File,Function Wrapper object use after free,Memory - illegal accesses,/agents/jvmpi/jvmpi_oprofile.cpp,compiled_method_load(JVMPI_Event *) Unchecked return value,Error handling issues,/daemon/opd_mangling.c,opd_open_sample_file Dereference after null check,Null pointer dereferences,/daemon/opd_sfile.c,sfile_hash Uninitialized scalar field,Uninitialized members,/gui/oprof_start_config.cpp,config_setting::config_setting() Division or modulo by zero,Integer handling issues,/libdb/db_stat.c,odb_hash_stat Resource leak,Resource leaks,/libop/op_cpu_type.c,_auxv_fetch Resource leak,Resource leaks,/libop/op_cpu_type.c,fetch_at_hw_platform Negative array index read,Memory - illegal accesses,/libop/op_events.c,_is_um_valid_bitmask Write to pointer after free,Memory - corruptions,/libop/op_events.c,read_events Read from pointer after free,Memory - illegal accesses</pre> Maynard Johnson 2013-05-20T22:41:54 http://comments.gmane.org/gmane.linux.oprofile/11505 <pre>The patch below is ppc64-specific functionality. As the ppc64 maintainer for oprofile, I have taken the liberty of committing it already. But as usual, I'd be happy to respond to any review comments. --------------------------------------------------------- Add support for architected events for IBM ppc64 architecture The Power ISA 2.07, recently published at http://power.org/, formally defines base performance monitoring facilities which must be provided by any processor implementation of the ISA. Specific implementations may provide additional features, but must include the standard architected features. This patch creates a generic ppc64 cpu type called "ppc64/architected_events_v1" that has a list of events which are defined in the ISA 2.07 performance monitoring unit architecture section. This new generic type will only be supported by operf. It will *not* be supported by the legacy oprofile kernel driver and opcontrol-based profiler. This new cpu type can be used in situations where oprofile i</pre> Maynard Johnson 2013-05-17T14:05:56 http://comments.gmane.org/gmane.linux.oprofile/11503 <pre>Hello, Giles, I don't have a Pentium P4 to test out this patch, so can you please apply it to the latest oprofile source and test it out. You can get the latest oprofile source from: git clone git://oprofile.git.sourceforge.net/gitroot/oprofile/oprofile Please compare results with the opcontrol-based profiler. Thanks for your patience. -Maynard ---------------------------------------------------- Add support for Intel Netburst (e.g., Pentium P4) to operf The "legacy" oprofile kernel driver has special "p4" handling. There's a map of event codes to ESCR/CCCR values. Unfortunately, the P4 event codes (stored in events/i386/p4/events) that are used by the oprofile kernel driver don't match what perf_events kernel code expects. This patch adds some p4-specific event code handling to operf in order to generate the correct encoding to pass to perf_events kernel subsystem. Signed-off-by: Maynard Johnson &lt;maynardj&lt; at &gt;us.ibm.com&gt; --- libop/Makefile.am | 4 +- libop/op_netburst.c | 1597 +++++++++++++</pre> Maynard Johnson 2013-05-15T15:35:05 http://comments.gmane.org/gmane.linux.oprofile/11497 <pre>I have registered the oprofile project with Coverity and requested a scan of the current source code in our git repo. I've not been able to figure out a way to generate the type of reports that Will has posted to the list in the past. Perhaps those types of reports are only available by using the Coverity product versus the free online version available to OSS projects. Below is a list of HIGH and MEDIUM impact issues the current scan found. This list also filters out STREAM_FORMAT_STATE errors, which I feel (as I've mentioned in the past) should be ignored. I've opened an oprofile bug report for these issues (https://sourceforge.net/tracker/?func=detail&amp;aid=3612911&amp;group_id=16191&amp;atid=116191). I would be happy to have a volunteer take ownership. -Maynard -------------------------------------------------------------------- Impact,Category,File,Function Medium,API usage errors,/pe_profiling/operf.cpp,_process_events_list() Medium,Error handling issues,/daemon/opd_mangling.c,opd_open_sample_file Me</pre> Maynard Johnson 2013-05-08T16:00:54 http://comments.gmane.org/gmane.linux.oprofile/11496 <pre>Bugs item #3612911, was opened at 2013-05-08 08:58 Message generated for change (Tracker Item Submitted) made by maynardj You can respond by visiting: https://sourceforge.net/tracker/?func=detail&amp;atid=116191&amp;aid=3612911&amp;group_id=16191 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Maynard Johnson (maynardj) Assigned to: Nobody/Anonymous (nobody) Summary: Coverity scan results: May 8, 2013 Initial Comment: I have registered the oprofile project with Coverity and requested a scan of the current source code in our git repo. I've not been able to figure out a way to generate the type of reports that Will Cohen has posted to the list in the past. Perhaps those types of reports are only available by using the Coverity product versus the free online version available to OSS projects. Below is a list o</pre> SourceForge.net 2013-05-08T15:58:39 http://comments.gmane.org/gmane.linux.oprofile/11495 <pre>This patch finally addresses the Coverity issues pointed out in the report posted by Will Cohen on Jan 15. I tested the changes by building the source code and running the oprofile-tests testsuite on various distros: - RHEL 6.4/Intel - RHEL 5.8/ppc64 - SLES 11 SP2/ppc64 - RHEL 6.3/ppc64 Advance Toolchain (gcc 4.7.3) -Maynard ------------------------------------------------------------------------ Fix Coverity issues identified against oprofile 0.9.8 release Signed-off-by: Maynard Johnson &lt;maynardj&lt; at &gt;us.ibm.com&gt; --- agents/jvmti/libjvmti_oprofile.c | 4 ++-- daemon/opd_anon.c | 2 +- daemon/opd_ibs.c | 7 ++++++- daemon/opd_mangling.c | 3 +++ daemon/opd_sfile.c | 1 + gui/oprof_start.cpp | 27 +++++++++++++++++++++++++-- gui/oprof_start_main.cpp | 12 +++++++++++- libop/op_mangle.c | 5 +++-- libopagent/jitdump.h | 10 ++++++++++ libopagent/opagent.c | 13 +</pre> Maynard Johnson 2013-05-08T15:37:57 http://comments.gmane.org/gmane.linux.oprofile/11493 <pre>Bugs item #3612843, was opened at 2013-05-07 11:37 Message generated for change (Tracker Item Submitted) made by maynardj You can respond by visiting: https://sourceforge.net/tracker/?func=detail&amp;atid=116191&amp;aid=3612843&amp;group_id=16191 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Maynard Johnson (maynardj) Assigned to: Maynard Johnson (maynardj) Summary: Feature request: oprofile should have an event counting tool Initial Comment: An event counting tool similar to 'perf stat' would be a nice addition to the oprofile toolbox. Using such a tool to do CPI analysis would result in more accurate results compared to a profiler, not to mention much less perturbation of the system (i.e., no PMU interrupts and saving sample information). Like operf, an oprofile event counting tool would understand symbo</pre> SourceForge.net 2013-05-07T18:37:46 http://comments.gmane.org/gmane.linux.oprofile/11492 <pre>Hi all, I'd like to propose the development of a new feature (or tool) for events counting so that some kind of performance analyze techniques, for instance, CPI (Cycles Per Instructions) breakdown, could take enormous benefits since they don't require a full application/system-wide profiling. Indeed, it is possible to leverage 'perf stat' for that purpose but that tool still lack some advantages present on Oprofile (e.g symbolic event names, XML output, post-processing tools) that really increase usability besides allowing high-level tools to be built on top of it. It seems now perfect timing to combine operf (kernel perf_events) with Oprofile post-processing/report tools to make a powerful counting tool. Is there any bugzilla feature request for that? Does anyone willing to work on? I appreciate any comments or suggestions on this topic. - Wainer ------------------------------------------------------------------------------ Learn Graph Databases - Download FREE O'Reilly Book "Graph Database</pre> Wainer Moschetta 2013-05-07T14:43:15 http://comments.gmane.org/gmane.linux.oprofile/11491 <pre>Hello. To analyze the behaviour of an application, I tried to use `operf' instead of `opcontrol' that is deprecated ( operf-0.9.8 ) Building Oprofile from source, with kernel headers package installed ( kernel 3.1.5 running on a P4-Celeron processor ) was perfectly right and `opcontrol' gave usable results for this application. But, in exactly the same conditions ( kernel-3.1.5, P4-Celeron processor, profiling for the default event ( GLOBAL_POWER_EVENTS ) ) `operf' failed to start. My analisys of `operf' setup shows that the system call to `perf_event_open` returns EINVAL; this return code comes from the kernel function `p4_validate_raw_event' that rejected the value of `attr.config' field in the event structure : the value provided by `operf' is 285 ( 1 &lt;&lt; 8 | 0x1d ) based on the code of GLOBAL_POWER_EVENTS ( 29 i.e. 0x1d ) according to the content of `/usr/share/events/i386/p4/events' There seems to be a mismatch between the value provided by operf and this expected by the kernel. Switching</pre> Gilles Allard 2013-05-07T10:20:48 http://comments.gmane.org/gmane.linux.oprofile/11471 <pre>Hi , I have a C file which calls a big loop function from a dynamically loaded shared library. When I run opreport I am not able to see the shared library in the report. My oprofile script looks like: opcontrol --reset opcontrol --setup --no-vmlinux opcontrol -start sleep 40 ./tpf sleep 40 opcontrol --stop opcontrol --dump opreport -lgt 0.1 --image-path=/lib/modules/2.6.16.60-0.97.TDC.1.R.0-smp/kernel --exclude-symbols mwait_idle,default_idle | /root/bin/tprof opreport -l -p /usr/local/bin/oprofiled --exclude-symbols mwait_idle,default_idle I know , I am missing something but I am not sure . Can anyone please help me to point out ? Thanks, Sandip ------------------------------------------------------------------------------ Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET Get 100% visibility into your production application - at no cost. Code-level diagnostics for performance bottlenecks with &lt;2% overhead Download for free and get started troubleshooting in minutes. http:</pre> Nandi, Sandip 2013-05-02T00:14:17 http://comments.gmane.org/gmane.linux.oprofile/11467 <pre>sir, i have installed oprofile-0.9.8 in my ubuntu 12.04 version and want to do code profiling i have tried operf ./fact opannotate --source ./fact but i am not getting annotated source code. help me in this regard thankyou navatha.k ------------------------------------------------------------------------------ Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET Get 100% visibility into your production application - at no cost. Code-level diagnostics for performance bottlenecks with &lt;2% overhead Download for free and get started troubleshooting in minutes. http://p.sf.net/sfu/appdyn_d2d_ap1_______________________________________________ oprofile-list mailing list oprofile-list&lt; at &gt;lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/oprofile-list </pre> Navatha D 2013-04-30T10:30:22 http://comments.gmane.org/gmane.linux.oprofile/11463 <pre>Hi Folks, I've trying to build OProfile from the latest source on a machine running SUSE Enterprise Desktop Linux 11 SP2 with a recent Intel core7 CPU and I'm encountering the following error: Making all in libperf_events make[2]: Entering directory `/local_static/git/oprofile/libperf_events' g++ -DHAVE_CONFIG_H -I. -I.. -I ../libabi -I ../libutil -I ../libutil++ -I ../libop -I ../libdb -I ../libperf_events -g -O2 -MT operf_utils.o -MD -MP -MF .deps/operf_utils.Tpo -c -o operf_utils.o operf_utils.cpp mv -f .deps/operf_utils.Tpo .deps/operf_utils.Po g++ -DHAVE_CONFIG_H -I. -I.. -I ../libabi -I ../libutil -I ../libutil++ -I ../libop -I ../libdb -I ../libperf_events -g -O2 -MT operf_counter.o -MD -MP -MF .deps/operf_counter.Tpo -c -o operf_counter.o operf_counter.cpp operf_counter.cpp: In constructor 'operf_counter::operf_counter(operf_event_t&amp;, bool, bool, bool, bool, int)': operf_counter.cpp:211: error: 'struct perf_event_attr' has no member named 'precise_ip' make[2]: *** [operf_counter.o] Error 1</pre> Langley, Scott E 2013-04-30T01:13:24 http://comments.gmane.org/gmane.linux.oprofile/11456 <pre>Hi, I tried to install oprofile-0.9.8 and got the below error. I was cross compiling the source for arm-none-linux-gnueabi compiler. I was not able to find the "iberty" library. Initially oi got error for "popt" library, which i installed. Please help. Cross compiling on Ubuntu 12.04 I used below command: user&lt; at &gt;ubu:~/oprofile-0.9.8$ CC=arm-none-linux-gnueabi-gcc ./configure --host=arm-none-linux-gnueabi --prefix=/home/user/tools --build=arm checking libiberty.h presence... no checking for libiberty.h... no checking for cplus_demangle in -liberty... no configure: error: liberty library not found user&lt; at &gt;ubu:~/oprofile-0.9.8$ Regards Ashoka. K ------------------------------------------------------------------------------ Try New Relic Now &amp; We'll Send You this Cool Shirt New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, &amp; servers with just a few lines of code. Try New Relic and get this </pre> Ashoka K 2013-04-27T13:25:35 http://comments.gmane.org/gmane.linux.oprofile/11454 <pre>This patch has already been pushed upstream, since it's simple and small and affects only the ppc64 architecture, which I maintain for oprofile. ------------------------------------------------------------------ Use PMC5/PMC6 on ppc64 arch for run cycles/run instructions The IBM Power processor architecture (ppc64) counts instructions and cycles on PMC5 and PMC6 (respectively) when the run latch is set (i.e., when not in idle state). On POWER6, these counters were not capable of generating interrupts, so they could not be used for profiling purposes; therefore, oprofile counted those events (PM_RUN_INST_CMPL and PM_RUN_CYC) using other counters. But with the newer POWER7 processor, PMC5 and PMC6 can generate interrupts, so it makes sense to leverage those two counters instead of using the other 4 (programmable) counters. Doing so could, theoretically allow us to count up to 6 events simultaneously without the kernel having to do multiplexing. This patch will force PM_RUN_INST_CMPL and PM_RUN_CYC to be </pre> Maynard Johnson 2013-04-26T19:23:09 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.linux.oprofile