- 25 Jan, 2013 2 commits
-
-
Jiri Olsa authored
Fixing the dynamic array format field parsing. Currently the event_read_fields function could segfault while parsing dynamic array other than string type. The reason is the event->pevent does not need to be set and gets dereferenced unconditionaly. Also adding proper initialization of field->elementsize based on the parsed dynamic type. Signed-off-by:
Jiri Olsa <jolsa@redhat.com> Acked-by:
Steven Rostedt <rostedt@goodmis.org> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> 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/1359060403-32422-1-git-send-email-jolsa@redhat.com [ committer note: Made a char pointer parameter const, as requested by Steven ] Signed-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
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 fixes from Arnaldo Carvalho de Melo: . Allow skipping problematic entries in 'perf test'. . Fix some namespace problems in the event parsing routines. . Add 'perf test' entry to make sure the python binding doesn't have linking problems. . Adjust 'perf test' attr tests verbosity levels. . Make tools/perf build with GNU make v3.80, fix from Al Cooper. . Do missing feature fallbacks in just one place, removing duplicated code in multiple tools. . Fix some memory leaks, from David Ahern. . Fix segfault when drawing out-of-bounds jumps, from Frederik Deweerdt. . Allow of casting an array of char to string in 'perf probe', from Hyeoncheol Lee. . Add support for wildcard in tracepoint system name, from Jiri Olsa. . Update FSF postal address to be URL's, from Jon Stanley. . Add anonymous huge page recognition, from Joshua Zhu. . Remove some needless feature test checks, from Namhyung Kim. . Multiple improvements to the sort routines, from Namhyung Kim. . Fix warning on '>=' operator in libtraceevent, from Namhyung Kim. . Use ARRAY_SIZE instead of reinventing it in 'perf script' and 'perf kmem', from Sasha Levin. . Remove some redundant checks, from Sasha Levin. . Test correct variable after allocation in libtraceevent, fix from Sasha Levin. . Mark branch_info maps as referenced, fix from Stephane Eranian. . Fix PMU format parsing test failure, from Sukadev Bhattiprolu. . Fix possible (unlikely) buffer overflow, from Thomas Jarosch. . Multiple 'perf script' fixes, from Tom Zanussi. . Add missing field in PERF_RECORD_SAMPLE documentation, from Vince Weaver. Signed-off-by:
Ingo Molnar <mingo@kernel.org>
-
- 24 Jan, 2013 38 commits
-
-
Arnaldo Carvalho de Melo authored
Sometimes a test is problematic for some reason and one wants to skip it, for instance: [root@sandy ~]# perf test 1: vmlinux symtab matches kallsyms : Ok 2: detect open syscall event : Ok 3: detect open syscall event on all cpus : Ok 4: read samples using the mmap interface : Ok 5: parse events tests : Warning: bad op token { Warning: bad op token { Warning: bad op token { Warning: bad op token { Warning: bad op token { Warning: function is_writable_pte not defined Segmentation fault (core dumped) So now we can use -s/--skip while the problematic tests are being fixed, allowing us to test all the other entries: [root@sandy ~]# perf test -s 5 1: vmlinux symtab matches kallsyms : Ok 2: detect open syscall event : Ok 3: detect open syscall event on all cpus : Ok 4: read samples using the mmap interface : Ok 5: parse events tests : Skip (user override) 6: x86 rdpmc test : Ok 7: Validate PERF_RECORD_* events & perf_sample fields : Ok 8: Test perf pmu format parsing : Ok 9: Test dso data interface : Ok 10: roundtrip evsel->name check : Ok 11: Check parsing of sched tracepoints fields : Ok 12: Generate and check syscalls:sys_enter_open event fields: Ok 13: struct perf_event_attr setup : Ok 14: Test matching and linking mutliple hists : Ok 15: Try 'use perf' in python, checking link problems : Ok [root@sandy ~]# 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@gmail.com> 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-klzd8p57jzdryafqkmlppcb1@git.kernel.orgSigned-off-by: -
Arnaldo Carvalho de Melo authored
Just like strlist allows passing a list of entries to parse. 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@gmail.com> 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-em50vqvvmlnc6k9tw4xtixus@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
So that we can work with optional parameters that may not set up an intlist. 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@gmail.com> 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-e9tmvgdzehqrza11zs0nbg7g@git.kernel.orgSigned-off-by:
-
Tom Zanussi authored
The tracepoints used by the workqueue-stats script no longer exist so trying to run the script results in: # perf script record workqueue-stats invalid or unsupported event: 'workqueue:workqueue_creation' Run 'perf list' for a list of valid events So remove the script until it can be reworked using the new workqueue tracepoints. Signed-off-by:
Tom Zanussi <tom.zanussi@linux.intel.com> Link: http://lkml.kernel.org/r/e7a7637d5df9df86887c3bff7683574665ec5360.1358527965.git.tom.zanussi@linux.intel.comSigned-off-by:
-
Tom Zanussi authored
Running the check-perf-trace scripts causes segfaults in both the Perl and Python cases: # perf script record check-perf-trace # perf script -s libexec/perf-core/scripts/python/check-perf-trace.py trace_begin Segmentation fault (core dumped) The reason is that the 'pevent' field was added to perf_scripting_context but it wasn't hooked up with an actual pevent in either case, so when one of the 'common' fields is accessed (in util/trace-event-parse.c:get_common_fields()), pevent->events tries to dereference a NULL pointer. This sets the pevent field when the scripting context is set up. Signed-off-by:
-
Tom Zanussi authored
Only display the trace info if using the default event display. When invoking scripts we assume they have complete control of what's displayed so we shouldn't unconditionally display the trace info, and when generating scripts we don't expect to see trace info obscuring the output message. Signed-off-by:
-
Tom Zanussi authored
For some reason the libtraceevent tracepoint-parsing code is missing the FIELD_IS_SIGNED flag-setting code, which causes problems for the Perl trace event binding at least, since it ends up unable to recognize negative numbers. Things like checking for negative return values therefore fail, causing scripts like rwtop to instead interpret the negative return value as a large positive value, which in turn get added to e.g. read totals with insanely invalid results. So set the FIELD_IS_SIGNED flag for tracepoint events that specify "signed:1". Before: # perf script record rw-by-pid # perf script report rw-by-pid read counts by pid: pid comm # reads bytes_requested bytes_read ------ -------------------- ----------- ---------- ---------- 753 Xorg 88 512000 7.74763251095801e+20 1619 firefox 42 462 2.58254417031934e+20 1232 gnome-shell 11 176 1.10680464442257e+20 1471 gnome-terminal 3 16366 18446744073709551615 1408 libsocialweb-co 2 32 18446744073709551613 After: # perf script report rw-by-pid read counts by pid: pid comm # reads bytes_requested bytes_read ------ -------------------- ----------- ---------- ---------- 753 Xorg 88 512000 2764 1619 firefox 42 462 126 1232 gnome-shell 11 176 40 1471 gnome-terminal 3 16366 10 1408 libsocialweb-co 2 32 8 Signed-off-by: -
Arnaldo Carvalho de Melo authored
Some would just call exit() anyway right after calling die() and the main routine doesn't have to call it, just return the exit value. 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@gmail.com> 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-nzq0sdur6oq6lgkt2ipf4o8s@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
They are only used in pmu.c, so no need to make them public in pmu.h. Acked-by:
-
Arnaldo Carvalho de Melo authored
We were using a homebrew equivalent, use the macro that is used everywhere for this function. Acked-by:
-
Arnaldo Carvalho de Melo authored
In tools/perf we use a convention where __ separates the struct name from the function name for functions that operate on a struct instance. Fix this usage by removing it from the struct names and fix also the associated functions. Acked-by:
-
Arnaldo Carvalho de Melo authored
In tools/perf we use a convention where __ separates the struct name from the function name for functions that operate on a struct instance. Fix this usage by removing it from the struct names and fix also the associated functions. Acked-by:
-
Arnaldo Carvalho de Melo authored
In tools/perf we use a convention where __ separates the struct name from the function name for functions that operate on a struct instance. Fix this usage by removing it from the struct names and fix also the associated functions. Acked-by:
-
Arnaldo Carvalho de Melo authored
In tools/perf we use a convention where __ separates the struct name from the function name for functions that operate on a struct instance. Fix this usage by removing it from the struct parse_events_term and fix also its associated functions. Acked-by:
-
Namhyung Kim authored
As we have ltrim() implementation in builtin-script.c move it to the more generic location of util/string.c so that it can be used from other places. Signed-off-by:
Namhyung Kim <namhyung@kernel.org> Cc: Feng Tang <feng.tang@intel.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1358845787-1350-14-git-send-email-namhyung@kernel.orgSigned-off-by:
-
Sukadev Bhattiprolu authored
On POWER, the 'perf format parsing' test always fails. Looks like it is because memset() is being passed number of longs rather than number of bytes. It is interesting that the test always passes on my x86 box. With this patch, the test passes on POWER and continues to pass on x86. Signed-off-by:
Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Acked-by:
-
Namhyung Kim authored
When make runs it tries to update the Makefile rules by reading all of included Makefiles. During the perf build it checks PERF-VERSION-FILE to get the current version number. But it triggers Makefile update so that make runs again with the update Makefile and, in turn, users will see duplicate CHK message on the second path. Running make with -d option for debugging tells me this: GNU Make 3.82 Built for x86_64-redhat-linux-gnu Copyright (C) 2010 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Reading makefiles... Reading makefile `Makefile'... Reading makefile `../scripts/Makefile.include' (search path) (no ~ expansion)... Reading makefile `config/utilities.mak' (search path) (no ~ expansion)... Reading makefile `PERF-VERSION-FILE' (search path) (don't care) (no ~ expansion)... Reading makefile `config/feature-tests.mak' (search path) (don't care) (no ~ expansion)... CHK -fstack-protector-all CHK -Wstack-protector CHK -Wvolatile-register-var ... Updating makefiles.... Considering target file `PERF-VERSION-FILE'. Must remake target `PERF-VERSION-FILE'. Invoking recipe from Makefile:52 to update target `PERF-VERSION-FILE'. Putting child 0x14037a0 (PERF-VERSION-FILE) PID 31925 on the chain. Live child 0x14037a0 (PERF-VERSION-FILE) PID 31925 PERF_VERSION = 3.8.rc3.gf751db6 Reaping winning child 0x14037a0 PID 31925 Removing child 0x14037a0 PID 31925 from chain. Successfully remade target file `PERF-VERSION-FILE'. ... Re-executing[1]: make -d <------------ here GNU Make 3.82 Built for x86_64-redhat-linux-gnu Copyright (C) 2010 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Reading makefiles... Reading makefile `Makefile'... Reading makefile `../scripts/Makefile.include' (search path) (no ~ expansion)... Reading makefile `config/utilities.mak' (search path) (no ~ expansion)... Reading makefile `PERF-VERSION-FILE' (search path) (don't care) (no ~ expansion)... Reading makefile `config/feature-tests.mak' (search path) (don't care) (no ~ expansion)... CHK -fstack-protector-all CHK -Wstack-protector CHK -Wvolatile-register-var ... Actually PERF-VERSION-FILE is used only for perf.c to #define PERF_VERSION macro. So make it like a C header file and include it during compiling the perf.c file will remove the need of being included into Makefile. Hench no need to update the Makefile and no CHK lines anymore. Signed-off-by:
-
Namhyung Kim authored
These lines are came from GIT Makefile and never used for perf. I found it from make -d output during working on previous patch. Updating makefiles.... Considering target file `arch/x86/Makefile'. No need to remake target `arch/x86/Makefile'. Considering target file `config.mak'. File `config.mak' does not exist. Must remake target `config.mak'. Failed to remake target file `config.mak'. Considering target file `config.mak.autogen'. File `config.mak.autogen' does not exist. Must remake target `config.mak.autogen'. Failed to remake target file `config.mak.autogen'. Signed-off-by:
-
Namhyung Kim authored
Although the '>=' (and '<=') operator is handled properly in libtraceevent, it emitted following spurious warnings on perf test: $ perf test 5: parse events tests : ... Warning: unknown op '>=' Warning: unknown op '>=' Warning: unknown op '>=' Warning: unknown op '>=' Warning: unknown op '>=' Warning: unknown op '>=' ... Add the operator to the checks. Signed-off-by:
-
Jiri Olsa authored
The 'unset' parameter is option callback leftover with no use, removing. Signed-off-by:
-
Stephane Eranian authored
As noticed by Jiri, the hist_entry->branch_info.to/from maps need to be marked as referenced to avoid problems later on. So we do this when the hist_entry is allocated. Signed-off-by:
Stephane Eranian <eranian@google.com> Acked-by:
-
Frederik Deweerdt authored
Factorize jump sanity checks from mark_jump_targets() and draw_current_jump() in an is_valid_jump() function. This fixes a segfault when moving the cursor over an invalid jump. Signed-off-by:
Frederik Deweerdt <frederik.deweerdt@xprog.eu> Cc: Namhyung Kim <namhyung@kernel.org> Link: http://lkml.kernel.org/r/20130114194716.GA4973@ks398093.ip-192-95-24.net [ committer note: Make it a disasm_line method ] Signed-off-by:
-
David Ahern authored
It is used by util/help.c so it should be a lib function and included in libperf.a. Code move only. Signed-off-by:
David Ahern <dsahern@gmail.com> Link: http://lkml.kernel.org/r/1358185681-90926-1-git-send-email-dsahern@gmail.comSigned-off-by:
-
David Ahern authored
Needs to be marked allocated so memory can be freed when dso is deleted. Signed-off-by:
-
David Ahern authored
Memory is currently leaked on some paths. Signed-off-by:
-
Namhyung Kim authored
Add description of sort keys to the perf-report document and also add missing cpu and srcline keys to the command line help string. Signed-off-by:
-
Namhyung Kim authored
Current perf report gets segmentation fault when a branch stack specific sort key is provided by --sort option to a perf.data file which contains no branch infomation. It's because those sort keys reference branch info of a hist entry unconditionally. Maybe we can change it checks whether such branch info is valid or not. But if the branch stacks are not recorded, it'd be nop. Thus it'd be better to make those keys are unselectable. This patch separates those keys to a different dimension array, so that if user passes such a key to a file which has no branch stack will get following message rather than a segfault. Error: Invalid --sort key: `symbol_from' Signed-off-by:
Stefan Beller <stefanbeller@googlemail.com> Acked-by:
-
Namhyung Kim authored
It doesn't need to compare to every sort key names since the index already has the required information. Signed-off-by:
-
Namhyung Kim authored
When hists__calc_col_len() called, most of column length are refreshed but it missed parent column. So if the parent sort key was used along with other keys rests will be misalinged since parent has no proper column width. Signed-off-by:
-
Namhyung Kim authored
Since cpu number is a natural number, it'd be more appropriate aligning it to right. Before: # Overhead CPU Command: Pid Shared Object # ........ ... ................. ..................... # 8.91% 8 gnome-shell: 1497 perf-1497.map 8.90% 7 gnome-shell: 1497 perf-1497.map 8.86% 9 gnome-shell: 1497 perf-1497.map 8.83% 6 gnome-shell: 1497 perf-1497.map 8.81% 10 gnome-shell: 1497 perf-1497.map 7.44% 5 gnome-shell: 1497 perf-1497.map 6.20% 3 gnome-shell: 1497 perf-1497.map 5.10% 0 gnome-shell: 1497 perf-1497.map After: # Overhead CPU Command: Pid Shared Object # ........ ... ................. ..................... # 8.91% 8 gnome-shell: 1497 perf-1497.map 8.90% 7 gnome-shell: 1497 perf-1497.map 8.86% 9 gnome-shell: 1497 perf-1497.map 8.83% 6 gnome-shell: 1497 perf-1497.map 8.81% 10 gnome-shell: 1497 perf-1497.map 7.44% 5 gnome-shell: 1497 perf-1497.map 6.20% 3 gnome-shell: 1497 perf-1497.map 5.10% 0 gnome-shell: 1497 perf-1497.map Signed-off-by: -
Namhyung Kim authored
The "pid" sort key prints "Command: Pid" output but it's misaligned. It's because of the offset of 6 was added to the column length during the calculation in order to reserve an space for Pid part but it isn't honored when printed. The output before this patch was like this: # Overhead Command: Pid Shared Object # ........ ............. ................. # 99.70% noploop:17814 noploop 0.29% noploop:17814 [kernel.kallsyms] 0.01% noploop:17814 ld-2.15.so Fix it by subtracting 6 for printing comm part. Signed-off-by: -
Namhyung Kim authored
Some functions have set __maybe_unused on its arguments that are used actually. Remove them. Signed-off-by:
-
Namhyung Kim authored
Some functions are misplaced along with other entries. Move them to a right place so that it can be found together with related functions. No functional change intended. Signed-off-by:
-
Hyeoncheol Lee authored
Before casting a type of a variable to string, convert_variable_type() confirms that the type is a pointer or an array. then if it is a pointer to char, it is casted to string. but in case of an array of char, it isn't Signed-off-by:
H.C. Lee <hyc.lee@gmail.com> Acked-by:
Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Link: http://lkml.kernel.org/r/CANFS6bb75e8a_UtyAD9yF73hfXDy0N8tSjDz=a+Vna=Y8ORMHg@mail.gmail.comSigned-off-by:
-
Vince Weaver authored
While trying to write a perf_event/mmap test for my perf_event test-suite I came across a missing field description in the PERF_RECORD_SAMPLE documentation in perf_event.h Signed-off-by:
Vince Weaver <vincent.weaver@maine.edu> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/alpine.DEB.2.02.1301081439300.24507@vincent-weaver-1.um.maine.eduSigned-off-by:
-
Joshua Zhu authored
Judging anonymous memory's vm_area_struct, perf_mmap_event's filename will be set to "//anon" indicating this vma belongs to anonymous memory. Once hugepage is used, vma's vm_file points to hugetlbfs. In this way, this vma will not be regarded as anonymous memory by is_anon_memory() in perf user space utility. Signed-off-by:
Joshua Zhu <zhu.wen-jie@hp.com> Cc: Akihiro Nagai <akihiro.nagai.hw@hitachi.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: David Ahern <dsahern@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Joshua Zhu <zhu.wen-jie@hp.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1357363797-3550-1-git-send-email-zhu.wen-jie@hp.comSigned-off-by:
-
Namhyung Kim authored
perf bench prints header message for bench suite before starting the benchmark. However if the stdout is redirected to a file and bench suite forks child processes this (and possibly other debugging messages too) will be repeated multiple times. $ perf bench sched messaging # Running sched/messaging benchmark... # 20 sender and receiver processes per group # 10 groups == 400 processes run Total time: 0.100 [sec] $ perf bench sched messaging > result.txt $ wc -l result.txt 391 In this file, there were so many "Running sched/messaging benchmark..." lines. This was because stdout is converted to fully-buffered due to the redirection and inherited child processes. Other lines are printed after reaping all those tasks. So fix it by flushing stdout before starting bench suites. Signed-off-by:Hitoshi Mitake <h.mitake@gmail.com> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> Cc: Ingo Molnar <mingo@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1357637966-8216-1-git-send-email-namhyung@kernel.orgSigned-off-by:
-
Namhyung Kim authored
The elf.h header file is used for NO_LIBELF build too so it should be included anyway. Also remove duplicated include of the header file in symbol-*.c. This patch fixes following build error on NO_LIBELF build: CC tests/hists_link.o tests/hists_link.c: In function ‘setup_fake_machine’: tests/hists_link.c:132:8: error: ‘STB_GLOBAL’ undeclared (first use in this function) tests/hists_link.c:132:8: note: each undeclared identifier is reported only once for each function it appears in Signed-off-by:
-
