- 17 Dec, 2015 40 commits
-
-
Jiri Olsa authored
Adding processing of event update events, so perf stat report can store additional info for events - unit,scale,name. Committer note: Before: # perf stat record -e power/energy-cores/ -a ^C Performance counter stats for 'system wide': 77.41 Joules power/energy-cores/ 1.597176695 seconds time elapsed # perf stat report Performance counter stats for '/home/acme/bin/perf stat record -e power/energy-cores/ -a': 332,488,114,176 power/energy-cores/ 1.597176695 seconds time elapsed # After, using the same perf.data file generated in the "Before" case above: # perf stat report Performance counter stats for '/home/acme/bin/perf stat record -e power/energy-cores/ -a': 77.41 Joules power/energy-cores/ 1.597176695 seconds time elapsed # Reported-by:Kan Liang <kan.liang@intel.com> Signed-off-by:
Jiri Olsa <jolsa@kernel.org> Tested-by:
Arnaldo Carvalho de Melo <acme@redhat.com> Cc: David Ahern <dsahern@gmail.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1446734469-11352-17-git-send-email-jolsa@kernel.orgSigned-off-by:
-
Jiri Olsa authored
Adding processing of stat and stat round events. The stat data com in stat events, using generic function process_stat_round_event to store data under perf_evsel object. The stat-round events comes each interval or as last event in non interval mode. The function process_stat_round_event process stored data for each perf_evsel object and print it out. Committer note: After this patch: $ perf stat record usleep 1 Performance counter stats for 'usleep 1': 0.498381 task-clock (msec) # 0.571 CPUs utilized 2 context-switches # 0.004 M/sec 0 cpu-migrations # 0.000 K/sec 149 page-faults # 0.299 M/sec 1,271,635 cycles # 2.552 GHz 928,712 stalled-cycles-frontend # 73.03% frontend cycles idle 663,286 stalled-cycles-backend # 52.16% backend cycles idle 792,614 instructions # 0.62 insns per cycle # 1.17 stalled cycles per insn 136,850 branches # 274.589 M/sec <not counted> branch-misses (0.00%) 0.000873419 seconds time elapsed $ $ perf stat report Performance counter stats for '/home/acme/bin/perf stat record usleep 1': 0.498381 task-clock (msec) # 0.571 CPUs utilized 2 context-switches # 0.004 M/sec 0 cpu-migrations # 0.000 K/sec 149 page-faults # 0.299 M/sec 1,271,635 cycles # 2.552 GHz 928,712 stalled-cycles-frontend # 73.03% frontend cycles idle 663,286 stalled-cycles-backend # 52.16% backend cycles idle 792,614 instructions # 0.62 insns per cycle # 1.17 stalled cycles per insn 136,850 branches # 274.589 M/sec <not counted> branch-misses (0.00%) 0.000873419 seconds time elapsed $ Reported-by: -
Jiri Olsa authored
So we have csv_sep properly initialized before report command leg. Reported-by:
-
Jiri Olsa authored
Using perf.data's perf_env data to initialize aggregate config. Reported-by:
-
Jiri Olsa authored
Adding processing of stat config event and initialize stat_config object. Reported-by:
-
Jiri Olsa authored
Adding processing of cpu/threads maps. Configuring session's evlist with these maps. Reported-by:
-
Jiri Olsa authored
Adding 'perf stat report' command support. ATM it only processes attr events and display nothing. Reported-by:
-
Jiri Olsa authored
Synthesize other events stuff not carried within attr event - unit, scale, name. Reported-by:
-
Jiri Olsa authored
We currently don't support storing multiple session in perf.data, so we can't allow -r option in stat record. $ perf stat -e cycles -r 2 record ls Cannot use -r option with perf stat record. Committer note: Before this patch we would a perf.data file such as: $ perf stat -e cycles -r 2 record ls <SNIP> Performance counter stats for 'ls' (2 runs): 3,935,236 cycles 0.002353261 seconds time elapsed ( +- 4.76% ) $ perf report -D | grep PERF_RECORD | grep ROUND 0xf0 [0]: failed to process type: 16 Error: failed to process sample $ Reported-by: -
Jiri Olsa authored
Writing stat round events on 'perf stat record' for each interval round. In non interval mode we store round event after the last stat event. Committer note: After the patch: $ perf report -D | grep PERF_RECORD | grep ROUND 0x852 [0x18]: PERF_RECORD_STAT_ROUND $ Reported-by:
-
Jiri Olsa authored
Writing stat events on 'perf stat record' at the time we read counter values from kernel. Committer note: After the patch: $ perf stat record usleep 1 Performance counter stats for 'usleep 1': 0.598006 task-clock (msec) # 0.484 CPUs utilized 1 context-switches # 0.002 M/sec 0 cpu-migrations # 0.000 K/sec 52 page-faults # 0.087 M/sec 882,744 cycles # 1.476 GHz 581,416 stalled-cycles-frontend # 65.86% frontend cycles idle <not supported> stalled-cycles-backend 636,479 instructions # 0.72 insns per cycle # 0.91 stalled cycles per insn 129,334 branches # 216.275 M/sec 7,512 branch-misses # 5.81% of all branches 0.001235157 seconds time elapsed $ oldperf evlist task-clock context-switches cpu-migrations page-faults cycles stalled-cycles-frontend stalled-cycles-backend instructions branches branch-misses $ oldperf report --stdio Error: The perf.data file has no samples! # To display the perf.data header info, please use --header/--header-only options. # $ perf report -D | grep PERF_RECORD 0x5b0 [0x28]: PERF_RECORD_THREAD_MAP nr: 1 thread: 5504 0x5d8 [0x12]: PERF_RECORD_CPU_MAP nr: 1 cpu: 65535 0x5ea [0x40]: PERF_RECORD_STAT_CONFIG 0x62a [0x30]: PERF_RECORD_STAT 0x65a [0x30]: PERF_RECORD_STAT 0x68a [0x30]: PERF_RECORD_STAT 0x6ba [0x30]: PERF_RECORD_STAT 0x6ea [0x30]: PERF_RECORD_STAT 0x71a [0x30]: PERF_RECORD_STAT 0x74a [0x30]: PERF_RECORD_STAT 0x77a [0x30]: PERF_RECORD_STAT 0x7aa [0x30]: PERF_RECORD_STAT -1 -1 0x7da [0x40]: PERF_RECORD_MMAP -1/0: [0xffffffff81000000(0x1f000000) @ 0xffffffff81000000]: x [kernel.kallsyms]_text $ Signed-off-by: -
Jiri Olsa authored
Allowing storing stat record data into pipe, so report tools (report/script) could read data directly from record. Committer note: Before this patch: $ perf stat record -o - usleep 1 | perf report -i - incompatible file format (rerun with -v to learn more) $ perf stat record -o - usleep 1 | perf script -i - incompatible file format (rerun with -v to learn more) $ ls -la perf.data ls: cannot access perf.data: No such file or directory $ After: $ perf stat record -o - usleep 1 | perf report -i - # To display the perf.data header info, please use # --header/--header-only options. # Error: The - file has no samples! $ perf stat record -o - usleep 1 | perf script -i - Display of symbols requested but neither sample IP nor sample address is selected. Hence, no addresses to convert to symbols. 0 [0x80]: failed to process type: 64 $ ls -la perf.data ls: cannot access perf.data: No such file or directory $ Signed-off-by:
-
Jiri Olsa authored
Store event IDs in evlist object so it get stored into perf.data file. Signed-off-by:
-
Jiri Olsa authored
Will be used to storing the event IDs in evlist object so it get stored into perf.data file. Signed-off-by:
-
Jiri Olsa authored
Synthesizing needed stat record data for report/script: - cpu/thread maps - stat config Committer note: New records generated on a perf.data file with this patch: $ perf report -D | grep PERF_RECORD_ 0x568 [0x28]: PERF_RECORD_THREAD_MAP nr: 1 thread: 29097 0x590 [0x12]: PERF_RECORD_CPU_MAP nr: 1 cpu: 65535 0x5a2 [0x40]: PERF_RECORD_STAT_CONFIG $ Signed-off-by:
-
Jiri Olsa authored
Disabling all non stat related features. Also as we now enable STAT feature in the data file, adding code to instruct session open to skip sample type checking for stat data files. Signed-off-by:
-
Jiri Olsa authored
Add 'perf stat record' command support. It creates simple (header only) perf.data file ATM. The record command could be specified anywhere among stat options. All stat command options are valid for stat record command with '-o' option exception. If specified for record command it denotes the perf data file name. Committer note: Set sample_type to PERF_SAMPLE_IDENTIFIER, which should be harmless while avoiding that older tools show confusing messages, for instance, with sample_type = 0, we get: $ perf stat record usleep 1 Performance counter stats for 'usleep 1': 0.630237 task-clock (msec) # 0.528 CPUs utilized 1 context-switches # 0.002 M/sec 0 cpu-migrations # 0.000 K/sec 52 page-faults # 0.083 M/sec 978,312 cycles # 1.552 GHz 671,931 stalled-cycles-frontend # 68.68% frontend cycles idle <not supported> stalled-cycles-backend 646,379 instructions # 0.66 insns per cycle # 1.04 stalled cycles per insn 131,046 branches # 207.931 M/sec 7,073 branch-misses # 5.40% of all branches 0.001193240 seconds time elapsed $ oldperf evlist WARNING: The perf.data file's data size field is 0 which is unexpected. Was the 'perf record' command properly terminated? non matching sample_type $ While with sample_type set to PERF_SAMPLE_IDENTIFIER, after we re-run 'perf stat record usleep' we get: $ oldperf evlist WARNING: The perf.data file's data size field is 0 which is unexpected. Was the 'perf record' command properly terminated? task-clock context-switches cpu-migrations page-faults cycles stalled-cycles-frontend stalled-cycles-backend instructions branches branch-misses $ Which at least shows the names of the events in the perf.data file. Additionally, such files, when passed to 'perf report' will produce: $ oldperf report --stdio WARNING: The perf.data file's data size field is 0 which is unexpected. Was the 'perf record' command properly terminated? Warning: Kernel address maps (/proc/{kallsyms,modules}) were restricted. Check /proc/sys/kernel/kptr_restrict before running 'perf record'. As no suitable kallsyms nor vmlinux was found, kernel samples can't be resolved. Samples in kernel modules can't be resolved as well. Error: The perf.data file has no samples! # To display the perf.data header info, please use --header/--header-only options. # $ Which is confusing and can be solved by just adding the kernel mmap record, which will also remove that warning about the data size field being equal to zero, after generating the mmap record: $ perf stat record usleep 1 Performance counter stats for 'usleep 1': 0.600796 task-clock (msec) # 0.478 CPUs utilized 1 context-switches # 0.002 M/sec 0 cpu-migrations # 0.000 K/sec 54 page-faults # 0.090 M/sec 886,844 cycles # 1.476 GHz 582,169 stalled-cycles-frontend # 65.65% frontend cycles idle <not supported> stalled-cycles-backend 638,344 instructions # 0.72 insns per cycle # 0.91 stalled cycles per insn 130,204 branches # 216.719 M/sec 7,500 branch-misses # 5.76% of all branches 0.001255897 seconds time elapsed $ oldperf evlist task-clock context-switches cpu-migrations page-faults cycles stalled-cycles-frontend stalled-cycles-backend instructions branches branch-misses $ oldperf report --stdio Error: The perf.data file has no samples! # To display the perf.data header info, please use --header/--header-only options. # [acme@zoo linux]$ No warnings, sensible output about what are the events in the perf.data file and also a "file has no samples" message, which indeed it doesn't. Signed-off-by: -
Jiri Olsa authored
Introducing the 'stat' feature to mark a perf.data as created by the 'perf stat record' command. It contains no data. It's needed so that the report tools (report/script) can differentiate sampling data from counting data, because they need to be treated in a different way. In the future it might be used to store the version of the stat storage system used. Signed-off-by:
-
Jiri Olsa authored
The 'perf report -D' command will now display detailed output for these newly added events: event_update thread_map cpu_map stat stat_config stat_round Signed-off-by:
-
Jiri Olsa authored
To display a 'event update' event for raw dump. Signed-off-by:
-
Jiri Olsa authored
Adding the cpumask 'event update' event, that stores/transfer the cpumask for a event. Signed-off-by:
-
Jiri Olsa authored
Adding name type 'event update' event, that stores/transfer events name. Event's name is stored within perf.data's EVENT_DESC feature, but we don't have it if we get the report data from pipe. Signed-off-by:
-
Jiri Olsa authored
A__allocdding scale type 'event update' event, that stores/transfer events scale value. The PMU events can define the scale value which is used to multiply events data. Signed-off-by:
-
Jiri Olsa authored
Adding unit type 'event update' event, that stores/transfer events unit name. The unit name is part of the perf stat output data. Signed-off-by:
-
Jiri Olsa authored
It'll serve as a base event for additional event attributes details, that are not part of the attr event. At the moment this event is just a dummy one without any specific functionality. The type value will distinguish the update event details. It'll come in the following patches. The idea for this event is to be extensible for any update that the event might need in the future. Signed-off-by:
-
Jiri Olsa authored
Introducing the following functions to display the stat events for raw dump. perf_event__fprintf_stat perf_event__fprintf_stat_round perf_event__fprintf_stat_config Signed-off-by:
-
Jiri Olsa authored
Introduce the perf_event__synthesize_stat_round function to synthesize a 'struct stat_round_event'. Signed-off-by:
-
Jiri Olsa authored
Adding the stat round event to be stored after each stat interval round, so that report tools (report/script) gets notified and process interval data. Signed-off-by:
-
Jiri Olsa authored
Introducing the perf_event__process_stat_event function to process a 'struct perf_stat' data from a stat event. Signed-off-by:
-
Jiri Olsa authored
Introduce the perf_event__synthesize_stat function to synthesize a 'struct stat_event'. Signed-off-by:
-
Jiri Olsa authored
Adding a stat event to store a 'struct perf_counter_values' for a given event/cpu/thread. Signed-off-by:
-
Jiri Olsa authored
Introducing the perf_event__read_stat_config function to read a struct perf_stat_config object data from a stat config event. Signed-off-by:
-
Jiri Olsa authored
Introduce the perf_event__synthesize_stat_config to synthesize a 'struct perf_stat_config'. Storing the stat config in the form of tag-value pairs will, I believe, sort out future version extensibility issues. Signed-off-by:
-
Jiri Olsa authored
Adding the stat config event to pass/store stat config data, so report tools (report/script) know how to interpret stat data. The config data is stored in a 'tag|value' way to allow for easy extension and backwards compatibility. Signed-off-by:
-
Jiri Olsa authored
To display a cpu_map event for raw dump. Signed-off-by:
-
Jiri Olsa authored
Introducing the cpu_map__new_event function to create a struct cpu_map object from a cpu_map event. Signed-off-by:
-
Jiri Olsa authored
Introduce the perf_event__synthesize_cpu_map function to synthesize a struct cpu_map. Added generic interface: cpu_map_data__alloc cpu_map_data__synthesize to make the cpu_map synthesizing usable for other events. Signed-off-by:
-
Jiri Olsa authored
Adding the cpu_map event to pass/store cpu maps as data in a pipe/perf.data. We store maps in 2 formats: - list of cpus - mask of cpus The format that takes less space is selected transparently in the following patch. The interface is made generic, so we could add the cpumap event data into another event in the following patches. Signed-off-by:
-
Jiri Olsa authored
To display a thread_map event for a raw dump. Signed-off-by:
-
Jiri Olsa authored
Introducing the thread_map__new_event function to create a struct thread_map object from a thread_map event. Signed-off-by:
-
