- 23 Dec, 2013 5 commits
-
-
Dongsheng Yang authored
When we synthesize an comm event, if machine is guest, we should use the pid of machine as the event->comm.pid, rather than tgid of thread. Signed-off-by:
Dongsheng Yang <yangds.fnst@cn.fujitsu.com> Cc: David Ahern <dsahern@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Namhyung Kim <namhyung.kim@lge.com> Link: http://lkml.kernel.org/r/22455abe107c618a361e7b667ad0f098f7c9b4a3.1387572416.git.yangds.fnst@cn.fujitsu.comSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
Dongsheng Yang authored
When we synthesize the mmap events of user space, if machine is guest, we should set the event->header.misc to PERF_RECORD_MISC_GUEST_USER, rather than PERF_RECORD_MISC_USER. Signed-off-by:
-
Dongsheng Yang authored
When we synthesize the threads, we are looking for the infomation under /proc. But it is only for host. This patch look for the path of proc under machine->root_dir, then XXX__synthesize_threads() functions can support guest machines. Signed-off-by:
-
Dongsheng Yang authored
This patch remove a TODO in thread__find_addr_map() and add support of PERF_RECORD_MISC_GUEST_USER. Signed-off-by:
-
Dongsheng Yang authored
Currently, if we use perf kvm --guestkallsyms --guestmodules report, we can not get the perf information from perf data file. All sample are shown as unknown. Reproducing steps: # perf kvm --guestkallsyms /tmp/kallsyms --guestmodules /tmp/modules record -a sleep 1 [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.624 MB perf.data.guest (~27260 samples) ] # perf kvm --guestkallsyms /tmp/kallsyms --guestmodules /tmp/modules report |grep % 100.00% [guest/6471] [unknown] [g] 0xffffffff8164f330 This bug was introduced by 207b5792 (perf kvm: Fix regression with guest machine creation). In original code, it uses perf_session__find_machine(), it means we deliver symbol to machine which has the same pid, if no machine found, deliver it to *default* guest. But if we use perf_session__findnew_machine() here, if no machine was found, new machine with pid will be built and added. Then the default guest which with pid == 0 will never get a symbol. And because the new machine initialized here has no kernel map created, the symbol delivered to it will be marked as "unknown". This patch here is to revert commit 207b5792 and fix the SEGFAULT bug in another way. Verification steps: # ./perf kvm --guestkallsyms /home/kallsyms --guestmodules /home/modules record -a sleep 1 [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.651 MB perf.data.guest (~28437 samples) ] # ./perf kvm --guestkallsyms /home/kallsyms --guestmodules /home/modules report |grep % 22.64% :6471 [guest.kernel.kallsyms] [g] update_rq_clock.part.70 19.99% :6471 [guest.kernel.kallsyms] [g] d_free 18.46% :6471 [guest.kernel.kallsyms] [g] bio_phys_segments 16.25% :6471 [guest.kernel.kallsyms] [g] dequeue_task 12.78% :6471 [guest.kernel.kallsyms] [g] __switch_to 7.91% :6471 [guest.kernel.kallsyms] [g] scheduler_tick 1.75% :6471 [guest.kernel.kallsyms] [g] native_apic_mem_write 0.21% :6471 [guest.kernel.kallsyms] [g] apic_timer_interrupt Signed-off-by:
David Ahern <dsahern@gmail.com> Cc: stable@vger.kernel.org # 3.3+ Cc: David Ahern <dsahern@gmail.com> Link: http://lkml.kernel.org/r/1387564907-3045-1-git-send-email-yangds.fnst@cn.fujitsu.comSigned-off-by:
-
- 20 Dec, 2013 2 commits
-
-
Namhyung Kim authored
Move those print functions under "if (use_browser == 0)" so that they don't interfere with TUI output. Maybe they can handle other UIs later. Signed-off-by:
Namhyung Kim <namhyung@kernel.org> Cc: David Ahern <dsahern@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1387516278-17024-3-git-send-email-namhyung@kernel.orgSigned-off-by:
-
Namhyung Kim authored
There're some places printing messages to stdout/err directly. It should be converted to use proper error printing functions instead. Signed-off-by:
-
- 19 Dec, 2013 28 commits
-
-
Arnaldo Carvalho de Melo authored
The addr_location struct should fully qualify an address, and to do that it should have in it the machine where the thread was found. Thus all functions that receive an addr_location now don't need to also receive a 'machine', those functions just need to access al->machine instead, just like it does with the other parts of an address location: al->thread, al->map, etc. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-o51iiee7vyq4r3k362uvuylg@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
The 'evsel' parameter is not used, ditch it, reducing the function signature. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-kx9temzdcy7mk2edya9c1tdu@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
Removing unused parameters. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-b7r7o80o2xwwtlzgqxv50foe@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
Removing unused parameters. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-fspmnjadohrik8uvhytyu8lp@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
The 'browser' arg _is_ used, so ditch the misplaced attribute. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-bo4dabkip5iikhk3x384ac46@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
Reduce typing, functions use class__method convention, so unlikely to clash with other libraries. This actually was discussed in the "Link:" referenced message below. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/r/20131112113427.GA4053@ghostprotocols.netSigned-off-by:
-
Jiri Olsa authored
Using global 'O' processing code because it's already setup due to the scripts/Makefile.include include. Using global variable OUTPUT instead of the local BUILD_OUTPUT. Signed-off-by:
Jiri Olsa <jolsa@redhat.com> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Steven Rostedt <rostedt@goodmis.org> Link: http://lkml.kernel.org/r/1387460527-15030-10-git-send-email-jolsa@redhat.comSigned-off-by:
-
Jiri Olsa authored
Using global QUIET_CLEAN build output variable and so we could have only single clean message: CLEAN libtraceevent Signed-off-by:
-
Jiri Olsa authored
Using global QUIET_INSTALL build output variable and factoring plugins installation so we could have only single install message for plugins: INSTALL trace_plugins Getting rid of local print_install. Signed-off-by:
-
Jiri Olsa authored
Using global QUIET_LINK build output variable and getting rid of local print_static_lib_build, print_plugin_build and print_shared_lib_compile. We no longer distinguish between shared and static library in the build message. It's differenced by the built file suffix, like: $ make ... LINK libtraceevent.a LINK libtraceevent.so Signed-off-by: -
Jiri Olsa authored
Adding global QUIET_CC_FPIC build output variable and getting rid of local print_fpic_compile and print_plugin_obj_compile. Signed-off-by:
-
Jiri Olsa authored
Using global QUIET_CC build output variable and getting rid of local print_compile. Signed-off-by:
-
Jiri Olsa authored
Removing print_app_build variable, because it's not needed. Signed-off-by:
-
Jiri Olsa authored
Moving QUIET_(CLEAN|INSTAL) variables into: tools/scripts/Makefile.include to be usable by other tools. The change to use them in libtraceevent is in following patches. Signed-off-by:
-
Jiri Olsa authored
Factoring make install tests to check for multiple files. Adding default set of installed files for install and install_bin tests. Putting the 'test' line into the log file instead to the screen as it gets more complex now. If the tests fails to find a file, following message is displayed: $ make -f tests/make make_install_bin - make_install_bin: cd . && make -f Makefile DESTDIR=/tmp/tmp.nCVuQoSHaJ install-bin failed to find: bin/perf Signed-off-by:
-
Arnaldo Carvalho de Melo authored
Reduce typing, functions use class__method convention, so unlikely to clash with other libraries. This actually was discussed in the "Link:" referenced message below. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/r/20131112113427.GA4053@ghostprotocols.netSigned-off-by:
-
Arnaldo Carvalho de Melo authored
Its a local struct and the functions use the __ separator from the class name to the method name, so its unlikely that this will clash with other namespaces. Save some typing then. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-r011tdv7ianars9jr9ur2n4q@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
1. Since all callers either test if it is less than zero or assign its result to an int variable, convert it from ssize_t to int; 2. There is just one use for the 'session' variable, so use rec->session directly instead; 3. No need to store the result of perf_data_file__write, since that result is either 'size' or -1, the later making the error result to be stored in 'errno' and accessed thru printf's %m in the pr_err call. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-xwsk964dp681fica3xlqhjin@git.kernel.orgSigned-off-by:
-
Jiri Olsa authored
Changing the file output code to use the newly added perf_data_file__write interface. No functional change intended. Signed-off-by: -
Jiri Olsa authored
Using the perf_data_file object to handle output file processing. No functional change intended. Signed-off-by:
-
Namhyung Kim authored
The va_end() in _eprintf() should be removed since the caller also invokes va_end(). Signed-off-by:
-
Arnaldo Carvalho de Melo authored
Continuing to try to remove the code duplication introduced with mem and branch hist entry code, this time providing prologue and epilogues to deal with callchains when processing samples. Acked-by:
-
Arnaldo Carvalho de Melo authored
Since it is now accessed just thru addr_map_symbol and hist_entry wrappers. Acked-by:
-
Arnaldo Carvalho de Melo authored
Those are just wrappers to annotation methods, so move them to annotate.c Acked-by:
-
Arnaldo Carvalho de Melo authored
Since it has a hist_entry, no need to skip the hist layer and use the underlying symbol one. Acked-by:
-
Arnaldo Carvalho de Melo authored
Since there are three calls that could receive just the struct addr_map_symbol pointer and call the symbol method. Acked-by:
-
Arnaldo Carvalho de Melo authored
Since now symbol__addr_inc_samples() does the auto alloc, no need to do it prior to calling hist_entry__inc_addr_samples. Acked-by:
-
Arnaldo Carvalho de Melo authored
Instead of open coding it in multiple places in 'report' and 'top'. Acked-by:
-
- 18 Dec, 2013 5 commits
-
-
Namhyung Kim authored
The pevent_filter_strerror() function is for receiving actual error message from pevent_errno value. To do that, add a static buffer to event_filter for saving internal error message If a failed function saved other information in the static buffer returns the information, otherwise returns generic error message. Signed-off-by:
-
Namhyung Kim authored
The @entry argument already has the info so no need to pass them. Signed-off-by:
-
Namhyung Kim authored
The commit 09600e0f ("perf tools: Compare dso's also when comparing symbols") added a comparison of dso when comparing symbol. But if the sort key already has dso, it doesn't need to do it again since entries have a different dso already filtered out. Signed-off-by:
-
Namhyung Kim authored
If a hist entry doesn't have symbol information, compare it with its address. Currently it only compares its level or whether it's NULL. This can lead to an undesired result like an overhead exceeds 100% especially when callchain accumulation is enabled by later patch. Signed-off-by:
-
Ingo Molnar authored
Merge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core Pull perf/core improvements and refactorings from Arnaldo Carvalho de Melo: New features: . In perf timechart: - Add backtrace support to CPU info . Print pid along the name . Add support for CPU topology . Add new option --highlight'ing threads, be it by name or, if a numeric value is provided, that run more than given duration. From Stanislav Fomichev. Refactorings: . Rename some struct DSO binary_type related members and methods, to clarify its purpose and need for differentiation from symtab_type, i.e. one is about the files .text, CFI, etc, i.e. its binary contents, and the other is about where the symbol table came from. . Convert to new topic libraries, starting with an API one (sysfs, debugfs, etc), renaming liblk in the process, from Borislav Petkov. . Get rid of some more panic() like error handling in libtraceevent, from Namhyung Kim. Signed-off-by:
Ingo Molnar <mingo@kernel.org>
-
