- 09 May, 2012 6 commits
-
-
David Ahern authored
Additional toggles have pushed the help line out of view on a modestly sized terminal (120 columns wide). Shorten it to just reminders. Signed-off-by:
David Ahern <dsahern@gmail.com> Link: http://lkml.kernel.org/r/1336510879-64610-1-git-send-email-dsahern@gmail.comSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
David Ahern authored
perf-record defaults to the H/W cycles event and if it is not supported falls back to cpu-clock. Reset the event name as well. Signed-off-by:
-
David Ahern authored
The 'perf top' command falls back to cpu-clock if the H/W cycles event is not supported, but the event name is not updated leading to a misleading header: PerfTop: 8 irqs/sec kernel:75.0% exact: 0.0% [1000Hz cycles], ... Update the event name when the event type is changed so that the header displays correctly: PerfTop: 794 irqs/sec kernel:100.0% exact: 0.0% [1000Hz cpu-clock], ... Signed-off-by:
-
David Ahern authored
perf stat on PPC currently fails to run: $ perf stat -- sleep 1 Error: open_counter returned with 6 (No such device or address). /bin/dmesg may provide additional information. Fatal: Not all events could be opened. The problem is that until 2.6.37 (behavior changed with commit b0a873eb) perf on PPC returns ENXIO when hw_perf_event_init() fails. With this patch we get the expected behavior: $ perf stat -v -- sleep 1 cycles event is not supported by the kernel. stalled-cycles-frontend event is not supported by the kernel. stalled-cycles-backend event is not supported by the kernel. instructions event is not supported by the kernel. branches event is not supported by the kernel. branch-misses event is not supported by the kernel. ... Signed-off-by:
-
David Ahern authored
perf-record on PPC is not falling back to cpu-clock: $ perf record -ag -fo /tmp/perf.data -- sleep 1 Error: sys_perf_event_open() syscall returned with 6 (No such device or address). /bin/dmesg may provide additional information. Fatal: No CONFIG_PERF_EVENTS=y kernel support configured? The problem is that until 2.6.37 (behavior changed with commit b0a873eb) perf on PPC returns ENXIO when hw_perf_event_init() fails. With this patch we get the expected behavior: $ perf record -ag -fo /tmp/perf.data -v -- sleep 1 Old kernel, cannot exclude guest or host samples. The cycles event is not supported, trying to fall back to cpu-clock-ticks [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.151 MB /tmp/perf.data (~6592 samples) ] Signed-off-by:
-
Jiri Olsa authored
Using PRIu64 for printing out u64 nr_events to fix compilation for x86 32 bits. Cc: Arun Sharma <asharma@fb.com> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com> Cc: Cyrill Gorcunov <gorcunov@openvz.org> Cc: Frank C. Eigler <fche@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Robert Richter <robert.richter@amd.com> Cc: Stephane Eranian <eranian@google.com> Cc: Tom Zanussi <tzanussi@gmail.com> Cc: Ulrich Drepper <drepper@gmail.com> Link: http://lkml.kernel.org/r/1335958638-5160-7-git-send-email-jolsa@redhat.comSigned-off-by:
Jiri Olsa <jolsa@redhat.com> Signed-off-by:
-
- 08 May, 2012 1 commit
-
-
Arnaldo Carvalho de Melo authored
Additionally we were not checking if a cpu list had been provided by the user. Fix that. Reported-by:
Namhyung Kim <namhyung@gmail.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.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-ao3zrouylwmt7h9ikj0krubi@git.kernel.orgSigned-off-by:
-
- 07 May, 2012 7 commits
-
-
Namhyung Kim authored
Use same function with perf record and top to share the code checks combinations of different switches. Signed-off-by:
Namhyung Kim <namhyung.kim@lge.com> Reviewed-by:
-
Namhyung Kim authored
There are places that check whether target task/cpu is given or not and some of them didn't check newly introduced uid or cpu list. Add and use three of helper functions to treat them properly. Signed-off-by:
-
Namhyung Kim authored
The perf_target__strerror() sets @buf to a string that describes the (perf_target-specific) error condition that is passed via @errnum. This is similar to strerror_r() and does same thing if @errnum has a standard errno value. Signed-off-by:
Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Reviewed-by:
-
Namhyung Kim authored
Add and use the modern perf_target__parse_uid() and get rid of the old parse_target_uid(). Signed-off-by:
-
Namhyung Kim authored
The perf_target_errno enumerations are used to indicate specific error cases on perf target operations. It'd help libperf being a more generic library. Signed-off-by:
-
Namhyung Kim authored
Currently, 'perf record -- sleep 1' creates a cpu map for all online cpus since it turns out calling cpu_map__new(NULL). Fix it. Also it is guaranteed that cpu_list is NULL if PID/TID is given by calling perf_target__validate(), so we can make the conditional bit simpler. This also fixes perf test 7 (Validate) failure on my 6 core machine: $ cat /sys/devices/system/cpu/online 0-11 $ ./perf test -v 7 7: Validate PERF_RECORD_* events & perf_sample fields: --- start --- perf_evlist__mmap: Operation not permitted ---- end ---- Validate PERF_RECORD_* events & perf_sample fields: FAILED! Signed-off-by:
-
Arnaldo Carvalho de Melo authored
Check if neither of --pid, --tid or --uid was specified and if so, set system_wide appropriately. Namhyung's patch would make using any of the above target specifiers emit a warning in perf_target__validate, since it would see target.system_wide set and one of the others as well. So set system_wide after validation. Suggested-by:
-
- 04 May, 2012 1 commit
-
-
Jiri Olsa authored
Currently if we cannot decide the size of the event, we guess next event possition by: "... check alignment, and increment a single u64 in the hope to catch on again 'soon'" This usually ends up with segfault or endless loop. It's better to admit the failure right away, then pretend nothing happened. It makes the life easier ;) Signed-off-by:
Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/20120416184251.GA11503@m.brq.redhat.comSigned-off-by:
-
- 02 May, 2012 14 commits
-
-
Namhyung Kim authored
If gtk2 support is not enabled (or failed for some reason) try TUI again instead of falling directly back to the stdio interface. Signed-off-by:
Pekka Enberg <penberg@kernel.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Namhyung Kim <namhyung@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Pekka Enberg <penberg@kernel.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1335761711-31403-6-git-send-email-namhyung.kim@lge.comSigned-off-by:
-
Namhyung Kim authored
Now setup_browser can handle gtk2 front-end so split the TUI code to ui/tui/setup.c in order to remove dependency. To this end, make ui__init/exit global symbols and take an argument. Also split gtk code to ui/gtk/setup.c. Signed-off-by:
-
Namhyung Kim authored
We use double underscore characters to distinguish its subsystem and actual function name. Signed-off-by:
-
Namhyung Kim authored
As perf doesn't allow to specify gtk command-line option, drop the arguments and pass NULL to gtk_init(). This makes the function easier to be called from setup_browser(). Signed-off-by:
-
Namhyung Kim authored
The setup_browser contained newt-related codes in it. As gtk front-end added recently, it should be more generic to handle both cases properly. So move newt codes to the ui__init() for now. Signed-off-by:
-
Namhyung Kim authored
For further work on perf_target, it'd be better off splitting the code into a separate file. Signed-off-by:
-
Namhyung Kim authored
There were some combinations of these switches that are not so appropriate IMHO. Since there are implicit priorities between them and they worked well anyway, but it ends up opening useless duplicated events. For example, 'perf stat -t <pid> -a' will open multiple events for the thread instead of one. Add explicit checks and warn user in perf_target__validate(). Signed-off-by:
-
Namhyung Kim authored
Now we have all information that needed to create cpu/thread maps in struct perf_target, it'd be better using it as an argument. Signed-off-by:
-
Namhyung Kim authored
The perf_target__validate function is used to check given PID/TID/UID/CPU target options and warn if some combination is impossible. Also this can make some arguments of parse_target_uid() function useless as it is checked before the call via our new helper. Signed-off-by:
-
Namhyung Kim authored
Use struct perf_target as it is introduced by previous patch. This is a preparation of further changes. Signed-off-by:
-
Namhyung Kim authored
Use struct perf_target as it is introduced by previous patch. This is a preparation of further changes. Signed-off-by:
-
Namhyung Kim authored
The perf_target struct will be used for taking care of cpu/thread maps based on user's input. Since it is used on various subcommands it'd better factoring it out. Thanks to Arnaldo for suggesting the better name. Signed-off-by:
-
Robert Richter authored
Include header fixes for ... bool: util/parse-events.h:31: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘have_tracepoints’ ... and types.h: util/parse-events.h:28: error: expected ‘)’ before ‘config’ util/parse-events.h:34: error: expected declaration specifiers or ‘...’ before ‘u64’ util/parse-events.h:45: error: expected ‘)’ before ‘type’ This happens if now other include files are included before util/parse-events.h. Signed-off-by:
Robert Richter <robert.richter@amd.com> Cc: Ingo Molnar <mingo@kernel.org> Link: http://lkml.kernel.org/r/1333643188-26895-2-git-send-email-robert.richter@amd.comSigned-off-by:
-
Arnaldo Carvalho de Melo authored
It was unconditionally printing debug stuff when in non -v mode we should just print the name and result of the test. Now: [root@sandy ~]# perf test rdpmc 6: x86 rdpmc test: Ok [root@sandy ~]# perf test -v rdpmc 6: x86 rdpmc test: --- start --- 0: 6030 1: 60030 2: 600050 3: 6000056 4: 60000070 5: 600000266 ---- end ---- x86 rdpmc test: Ok [root@sandy ~]# Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.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-0tjedaozsy9oarq30nvzg74b@git.kernel.orgSigned-off-by:
-
- 23 Apr, 2012 1 commit
-
-
Ulrich Drepper authored
The if branch is completely unnecessary since 'realloc' handles NULL pointers for the first parameter. This is really only a cleanup and submitted mainly to prevent proliferation of bad practices. Signed-off-by:
Ulrich Drepper <drepper@gmail.com> Cc: David Ahern <dsahern@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Robert Richter <robert.richter@amd.com> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/r/201204231304.q3ND4TFe020805@drepperk.user.openhosting.comSigned-off-by:
-
- 13 Apr, 2012 2 commits
-
-
Ingo Molnar authored
Merge Linux 3.4-rc2: we were on v3.3, update the base. Signed-off-by:Ingo Molnar <mingo@kernel.org>
-
Ingo Molnar authored
Merge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core Fixes and improvements for perf/core: . Overhaul the tools/ makefiles, gluing them to the top level Makefile, from Borislav Petkov. . Move the UI files from tools/perf/util/ui/ to tools/perf/ui/. Also move the GTK+ browser to tools/perf/ui/gtk/, from Namhyung Kim. . Only fallback to sw cycles counter on ENOENT for the hw cycles, from Robert Richter . Trivial fixes from Robert Richter . Handle the autogenerated bison/flex files better, from Namhyung and Jiri Olsa. . Navigate jump instructions in the annotate browser, just press enter or ->, still needs support for a jump navigation history, i.e. to go back. . Search string in the annotate browser: same keys as vim: / forward n next backward/forward ? backward . Clarify number of events/samples in the report header, from Ashay Rane Signed-off-by:
-
- 11 Apr, 2012 8 commits
-
-
Robert Richter authored
Use cpu-clock-tick sw counter for cpu-cycles only if there is no hw pmu available. This is the case if the syscall reports ENOENT. In other cases (e.g. invalid attributes) we don't want the sw counter to be used. Cc: Ingo Molnar <mingo@kernel.org> Link: http://lkml.kernel.org/r/1333643188-26895-5-git-send-email-robert.richter@amd.comSigned-off-by:
-
Robert Richter authored
Thread map is actually type pid_t and not int. Signed-off-by:
-
Robert Richter authored
This references are not exported, use static declaration. Signed-off-by:
-
Namhyung Kim authored
Move those files to new directory in order to be prepared to further UI work. Makefile and header file pathes are adjusted accordingly. Also fix a build breakage if NO_GTK2=1 is given. Signed-off-by:
-
Namhyung Kim authored
Move those files to new directory in order to be prepared to further UI work. Makefile and header file pathes are adjusted accordingly. Signed-off-by:
-
Namhyung Kim authored
CC util/annotate.o util/annotate.c: In function symbol__annotate: util/annotate.c:87:16: error: parsed_line may be used uninitialized in this function [-Werror=maybe-uninitialized] util/annotate.c:211:22: note: parsed_line was declared here cc1: all warnings being treated as errors make: *** [util/annotate.o] Error 1 make: *** Waiting for unfinished jobs.... Signed-off-by: -
Jiri Olsa authored
Currently the parsers objects (bison/flex related) are each time perf is built. No matter the generated files are already in place, the parser generation is executed every time. Changing the rules to have proper flex/bison objects generation dependencies. The parsers code is not rebuilt until the flex/bison source files are touched. Also when flex/bison source is changed, only dependent objects are rebuilt. Signed-off-by:
-
Borislav Petkov authored
Now you can do $ make tools/<toolname> from the toplevel kernel directory and have the respective tool built. If you want to build and install it, do $ make tools/<toolname>_install $ make tools/<toolname>_clean should clean the respective tool directories. If you want to clean all in tools, simply do $ make tools/clean Also, if you want to get what the possible targets are, simply calling $ make tools/ should give you the short help. $ make tools/install installs all tools, of course. Doh. Signed-off-by:
Borislav Petkov <borislav.petkov@amd.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Michal Marek <mmarek@suse.cz> Cc: Sam Ravnborg <sam@ravnborg.org> Link: http://lkml.kernel.org/r/1334162178-17152-6-git-send-email-bp@amd64.orgSigned-off-by:
-
