Re: [GIT PULL 00/59] perf/core improvements and fixes

From: Ingo Molnar
Date: Sat Jan 09 2016 - 11:20:17 EST



* Arnaldo Carvalho de Melo <acme@xxxxxxxxxx> wrote:

> Hi Ingo,
>
> Please consider pulling, more to come, bpf, make build-test and
> hopefully Stephane's JIT stuff, etc.
>
> Best Regards,
>
> - Arnaldo
>
> The following changes since commit 9cc2617de5b9222abb39cd02e90d57dfea99c6d7:
>
> perf/x86/amd: Remove l1-dcache-stores event for AMD (2016-01-06 11:15:39 +0100)
>
> are available in the git repository at:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git tags/perf-core-for-mingo
>
> for you to fetch changes up to 775d8a1b0d75211cc6123915c6b5b688f2002478:
>
> perf evlist: Add --trace-fields option to show trace fields (2016-01-08 14:23:02 -0300)
>
> ----------------------------------------------------------------
> perf/core improvements and fixes:
>
> New features:
>
> - Allow using trace events fields as sort order keys, making 'perf evlist --trace_fields'
> show those, and then the user can select a subset and use like:
>
> perf top -e sched:sched_switch -s prev_comm,next_comm
>
> That works as well in 'perf report' when handling files containing
> tracepoints.
>
> The default when just tracepoint events are found in a perf.data file is to
> format it like ftrace, using the libtraceevent formatters, plugins, etc (Namhyung Kim)
>
> - Add support in 'perf script' to process 'perf stat record' generated files,
> culminating in a python perf script that calculates CPI (Cycles per
> Instruction) (Jiri Olsa)
>
> - Show random perf tool tips in the 'perf report' bottom line (Namhyung Kim)
>
> - perf report now defaults to --group if the perf.data file has grouped events, try it with:
>
> # perf record -e '{cycles,instructions}' -a sleep 1
> [ perf record: Woken up 1 times to write data ]
> [ perf record: Captured and wrote 1.093 MB perf.data (1247 samples) ]
> # perf report
> # Samples: 1K of event 'anon group { cycles, instructions }'
> # Event count (approx.): 1955219195
> #
> # Overhead Command Shared Object Symbol
>
> 2.86% 0.22% swapper [kernel.kallsyms] [k] intel_idle
> 1.05% 0.33% firefox libxul.so [.] js::SetObjectElement
> 1.05% 0.00% kworker/0:3 [kernel.kallsyms] [k] gen6_ring_get_seqno
> 0.88% 0.17% chrome chrome [.] 0x0000000000ee27ab
> 0.65% 0.86% firefox libxul.so [.] js::ValueToId<(js::AllowGC)1>
> 0.64% 0.23% JS Helper libxul.so [.] js::SplayTree<js::jit::LiveRange*, js::jit::LiveRange>::splay
> 0.62% 1.27% firefox libxul.so [.] js::GetIterator
> 0.61% 1.74% firefox libxul.so [.] js::NativeSetProperty
> 0.61% 0.31% firefox libxul.so [.] js::SetPropertyByDefining
>
> User visible fixes:
>
> - Coect data mmaps so that the DWARF unwinder can handle usecases needing them,
> like softice (Jiri Olsa)
>
> - Decay callchains in fractal mode, fixing up cases where 'perf top -g' would
> show entries with more than 100% (Namhyung Kim)
>
> Infrastructure:
>
> - Sync tools/lib with the lib/ in the kernel sources for find_bit.c and
> move bitmap.[ch] from tools/perf/util/ to tools/lib/ (Arnaldo Carvalho de Melo)
>
> - No need to set attr.sample_freq in some 'perf test' entries that only
> want to deal with PERF_RECORD_ meta-events, improve a bit error output
> for CQM test (Arnaldo Carvalho de Melo)
>
> - Fix python binding build, adding some missing object files now required
> due to cpumap using find_bit stuff (Arnaldo Carvalho de Melo)
>
> - tools/build improvemnts (Jiri Olsa)
>
> - Add more files to cscope/ctags databases (Jiri Olsa)
>
> - Do not show 'trace' in 'perf help' if it is not compiled in (Jiri Olsa)
>
> - Make perf_evlist__open() open evsels with their cpus and threads,
> like perf record does, making them consistent (Adrian Hunter)
>
> - Fix pmu snapshot initialization bug (Stephane Eranian)
>
> - Add missing headers in perf's MANIFEST (Wang Nan)
>
> Signed-off-by: Arnaldo Carvalho de Melo <acme@xxxxxxxxxx>
>
> ----------------------------------------------------------------
> Adrian Hunter (1):
> perf evlist: Make perf_evlist__open() open evsels with their cpus and threads (like perf record does)
>
> Arnaldo Carvalho de Melo (10):
> perf tests: No need to set attr.sample_freq in the perf time to TSC test
> perf evlist: Introduce perf_evlist__new_dummy constructor
> perf test: Use "dummy" events in the PERF_RECORD_ test
> perf test: No need for setting attr.sample_freq on the RECORD test
> perf python: Add missing files to binding link list
> perf tests: No need to set attr.sample_freq for tracking !PERF_RECORD_SAMPLE
> perf tests: Give a bit more information on the CQM test failure path
> tools lib: Move find_next_bit.c to tools/lib/
> tools lib: Sync tools/lib/find_bit.c with the kernel
> tools lib: Move bitmap.[ch] from tools/perf/ to tools/{lib,include}/
>
> Jiri Olsa (25):
> tools build feature: Fix feature_check_display_code typo
> tools build feature: Move dwarf post unwind choice output into perf
> tools build feature: Introduce feature_assign macro
> tools build feature: Use value assignment form for FEATURE-DUMP file
> perf build: Use FEATURE-DUMP in bpf subproject
> perf stat record: Keep sample_type 0 for pipe session
> perf script: Process cpu/threads maps
> perf script: Process stat config event
> perf script: Add process_stat/process_stat_interval scripting interface
> perf script: Add stat default handlers
> perf script: Add python support for stat events
> perf cpumap: Fix cpu conversion in cpu_map__from_entries
> perf script: Display stat events by default
> perf script: Add stat-cpi.py script
> perf tools: Do not show trace command if it's not compiled in
> perf script: Align event name properly
> perf tools: Include all tools/lib directory for tags/cscope/TAGS targets
> perf tools: Remove list entry from struct sort_entry
> perf tools: Add overhead/overhead_children keys defaults via string
> perf diff: Use perf_hpp__register_sort_field interface
> perf evlist: Remove perf_evlist__(enable|disable)_event functions
> perf unwind: Use find_map function in access_dso_mem
> perf unwind: Check for mmaps also in MAP__VARIABLE tree
> perf libdw: Check for mmaps also in MAP__VARIABLE tree
> perf record: Store data mmaps for dwarf unwind
>
> Namhyung Kim (21):
> perf hist: Pass struct sample to __hists__add_entry()
> perf hist: Save raw_data/size for tracepoint events
> tools lib traceevent: Factor out and export print_event_field[s]()
> perf top: Create the evlist sooner
> perf tools: Pass evlist to setup_sorting()
> perf tools: Add dynamic sort key for tracepoint events
> perf tools: Try to show pretty printed output for dynamic sort keys
> perf tools: Add 'trace' sort key
> perf report/top: Add --raw-trace option
> perf tools: Support shortcuts for events in dynamic sort keys
> perf tools: Support '<event>.*' dynamic sort key
> perf tools: Skip dynamic fields not defined for current event
> perf tools: Add 'trace_fields' dynamic sort key
> perf tools: Make 'trace' or 'trace_fields' sort key default for tracepoint events
> perf tools: Add all matching dynamic sort keys for field name
> perf report: Add documentation for dynamic sort keys
> perf top: Decay periods in callchains
> perf report: Change default to use event group view
> perf hists: Export a couple of hist functions
> perf report: Show random usage tip on the help line
> perf evlist: Add --trace-fields option to show trace fields
>
> Stephane Eranian (1):
> perf pmu: fix alias->snapshot missing initialization bug
>
> Wang Nan (1):
> perf tools: Add missing headers in perf's MANIFEST
>
> tools/build/Makefile.feature | 41 +-
> tools/{perf/util => }/include/linux/bitmap.h | 2 +
> tools/{perf/util => lib}/bitmap.c | 0
> tools/lib/bpf/Makefile | 4 +
> tools/lib/find_bit.c | 84 +++
> tools/lib/traceevent/event-parse.c | 129 ++---
> tools/lib/traceevent/event-parse.h | 4 +
> tools/lib/util/find_next_bit.c | 89 ---
> tools/perf/Build | 1 +
> tools/perf/Documentation/perf-evlist.txt | 3 +
> tools/perf/Documentation/perf-report.txt | 27 +
> tools/perf/Documentation/perf-top.txt | 3 +
> tools/perf/Documentation/tips.txt | 14 +
> tools/perf/MANIFEST | 4 +-
> tools/perf/Makefile.perf | 7 +-
> tools/perf/arch/x86/tests/intel-cqm.c | 2 +-
> tools/perf/arch/x86/tests/perf-time-to-tsc.c | 1 -
> tools/perf/arch/x86/util/intel-bts.c | 4 +-
> tools/perf/arch/x86/util/intel-pt.c | 4 +-
> tools/perf/builtin-annotate.c | 9 +-
> tools/perf/builtin-diff.c | 15 +-
> tools/perf/builtin-evlist.c | 11 +-
> tools/perf/builtin-record.c | 6 +-
> tools/perf/builtin-report.c | 6 +-
> tools/perf/builtin-script.c | 170 +++++-
> tools/perf/builtin-stat.c | 9 +-
> tools/perf/builtin-top.c | 16 +-
> tools/perf/command-list.txt | 2 +-
> tools/perf/config/Makefile | 10 +
> tools/perf/perf.c | 4 +
> tools/perf/scripts/python/stat-cpi.py | 77 +++
> tools/perf/tests/hists_cumulate.c | 8 +-
> tools/perf/tests/hists_filter.c | 2 +-
> tools/perf/tests/hists_link.c | 8 +-
> tools/perf/tests/hists_output.c | 10 +-
> tools/perf/tests/keep-tracking.c | 3 +-
> tools/perf/tests/perf-record.c | 6 +-
> tools/perf/tests/switch-tracking.c | 6 +-
> tools/perf/ui/browsers/hists.c | 7 +-
> tools/perf/ui/gtk/hists.c | 4 +-
> tools/perf/ui/hist.c | 14 +-
> tools/perf/ui/stdio/hist.c | 6 +-
> tools/perf/util/Build | 11 +-
> tools/perf/util/callchain.c | 28 +
> tools/perf/util/callchain.h | 1 +
> tools/perf/util/cpumap.c | 13 +-
> tools/perf/util/evlist.c | 72 +--
> tools/perf/util/evlist.h | 7 +-
> tools/perf/util/evsel.c | 23 +
> tools/perf/util/evsel.h | 1 +
> tools/perf/util/generate-cmdlist.sh | 15 +
> tools/perf/util/hist.c | 78 ++-
> tools/perf/util/hist.h | 24 +-
> tools/perf/util/pmu.c | 1 +
> tools/perf/util/python-ext-sources | 2 +
> .../util/scripting-engines/trace-event-python.c | 115 +++-
> tools/perf/util/sort.c | 601 ++++++++++++++++++++-
> tools/perf/util/sort.h | 12 +-
> tools/perf/util/symbol.c | 1 +
> tools/perf/util/symbol.h | 3 +-
> tools/perf/util/trace-event.h | 4 +
> tools/perf/util/unwind-libdw.c | 10 +
> tools/perf/util/unwind-libunwind.c | 20 +-
> tools/perf/util/util.c | 27 +
> tools/perf/util/util.h | 2 +
> 65 files changed, 1556 insertions(+), 347 deletions(-)
> rename tools/{perf/util => }/include/linux/bitmap.h (95%)
> rename tools/{perf/util => lib}/bitmap.c (100%)
> create mode 100644 tools/lib/find_bit.c
> delete mode 100644 tools/lib/util/find_next_bit.c
> create mode 100644 tools/perf/Documentation/tips.txt
> create mode 100644 tools/perf/scripts/python/stat-cpi.py

Pulled, thanks a lot Arnaldo!

Ingo