Skip to contentSkip to conversation

perf/core improvements and fixes

GIT PULL In review Arnaldo Carvalho de Melo 3 participants 20 messages Feb 21, 2019 – Feb 28, 2019 perf events

lore
Share & tools
lore

Arnaldo Carvalho de Melo

author maintainer
#0 lore raw

Hi Ingo,

Please consider pulling,

- Arnaldo

Test results at the end of this message, as usual.

The following changes since commit 43f4e6279f05eefac058a3524e184cecae463bfe:

Merge tag 'perf-core-for-mingo-5.1-20190214' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core (2019-02-15 10:19:11 +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-5.1-20190220

for you to fetch changes up to b4409ae112caa6315f6ee678e953b9fc93e6919c:

perf tools: Make rm_rf() remove single file (2019-02-20 17:09:28 -0300)

----------------------------------------------------------------
perf/core improvements and fixes:

perf report:

He Kuang:

- Don't shadow inlined symbol with different addr range.

perf script:

Jiri Olsa:

- Allow +- operator to ask for -F to add/remove fields to the default set, for instance to ask for the removal of the 'cpu' field in tracepoint events, adding 'period' to that kind of events, etc.

perf test:

Thomas Richter:

- Fix scheduler tracepoint signedness of COMM fields failure of 'evsel-tp-sched' test on s390 and other arches.

Tommi Rantala:

- Skip trace+probe_vfs_getname.sh when 'perf trace' is not built.

perf trace:

Arnaldo Carvalho de Melo:

- Add initial BPF map dumper, initially just for the current, minimal needs of the augmented_raw_syscalls BPF example used to collect pointer args payloads that uses BPF maps for pid and syscall filtering, but will in time have features similar to 'perf stat' --interval-print, --interval-clear, ways to signal from a BPF event that a specific map (or range of that map) should be printed, optionally as a histogram, etc.

General:

Jiri Olsa:

- Add cpu and numa topologies classes for further reuse, fixing some issues in the process.

- Fixup some warnings and debug levels.

- Make rm_rf() remove single file, not just directories.

Documentation:

Jonas Rabenstein:

- Fix HEADER_CMDLINE description in perf.data documentation.

- Fix documentation of the Flags section in perf.data.

----------------------------------------------------------------
Arnaldo Carvalho de Melo (2):
perf bpf: Add bpf_map dumper
perf trace: Allow dumping a BPF map after setting up BPF events

He Kuang (1):
perf report: Don't shadow inlined symbol with different addr range

Jiri Olsa (10):
perf header: Fix wrong node write in NUMA_TOPOLOGY feature perf tools: Add cpu_topology object
perf tools: Add numa_topology object
perf tools: Use sysfs__mountpoint() when reading cpu topology perf session: Don't report zero period samples for slave events perf evsel: Force sample_type for slave events
perf script: Allow +- operator for type specific fields option perf bpf-event: Add missing new line into pr_debug call perf cpumap: Increase debug level for cpu_map__snprint verbose output perf tools: Make rm_rf() remove single file

Jonas Rabenstein (2):
perf doc: Fix HEADER_CMDLINE description in perf.data documentation perf doc: Fix documentation of the Flags section in perf.data

Thomas Richter (1):
perf test: Fix failure of 'evsel-tp-sched' test on s390

Tommi Rantala (1):
perf tests shell: Skip trace+probe_vfs_getname.sh if built without trace support

 tools/perf/Documentation/perf-script.txt           |   6 +
 tools/perf/Documentation/perf-trace.txt            |   8 +
 tools/perf/Documentation/perf.data-file-format.txt |  11 +-
 tools/perf/builtin-script.c                        |   8 +
 tools/perf/builtin-trace.c                         |  19 ++
 tools/perf/tests/evsel-tp-sched.c                  |   6 +-
 tools/perf/tests/shell/lib/probe.sh                |   5 +
 tools/perf/tests/shell/trace+probe_vfs_getname.sh  |   1 +
 tools/perf/util/Build                              |   2 +
 tools/perf/util/bpf-event.c                        |   2 +-
 tools/perf/util/bpf_map.c                          |  72 ++++++
 tools/perf/util/bpf_map.h                          |  22 ++
 tools/perf/util/cpumap.c                           |   2 +-
 tools/perf/util/cputopo.c                          | 277 +++++++++++++++++++++
 tools/perf/util/cputopo.h                          |  33 +++
 tools/perf/util/evsel.c                            |   8 +
 tools/perf/util/header.c                           | 269 +++-----------------
 tools/perf/util/session.c                          |   7 +
 tools/perf/util/sort.c                             |  10 +-
 tools/perf/util/srcline.c                          |   2 +-
 tools/perf/util/util.c                             |  16 +-
 21 files changed, 530 insertions(+), 256 deletions(-)
 create mode 100644 tools/perf/util/bpf_map.c
 create mode 100644 tools/perf/util/bpf_map.h
 create mode 100644 tools/perf/util/cputopo.c
 create mode 100644 tools/perf/util/cputopo.h

Test results:

The first ones are container based builds of tools/perf with and without libelf support. Where clang is available, it is also used to build perf with/without libelf, and building with LIBCLANGLLVM=1 (built-in clang) with gcc and clang when clang and its devel libraries are installed.

The objtool and samples/bpf/ builds are disabled now that I'm switching from using the sources in a local volume to fetching them from a http server to build it inside the container, to make it easier to build in a container cluster. Those will come back later.

Several are cross builds, the ones with -x-ARCH and the android one, and those may not have all the features built, due to lack of multi-arch devel packages, available and being used so far on just a few, like debian:experimental-x-{arm64,mipsel}.

The 'perf test' one will perform a variety of tests exercising tools/perf/util/, tools/lib/{bpf,traceevent,etc}, as well as run perf commands with a variety of command line event specifications to then intercept the sys_perf_event syscall to check that the perf_event_attr fields are set up as expected, among a variety of other unit tests.

Then there is the 'make -C tools/perf build-test' ones, that build tools/perf/ with a variety of feature sets, exercising the build with an incomplete set of features as well as with a complete one. It is planned to have it run on each of the containers mentioned above, using some container orchestration infrastructure. Get in contact if interested in helping having this in place.

  # export PERF_TARBALL=http://192.168.124.1/perf/perf-5.0.0-rc5.tar.xz
  # dm
   1 alpine:3.4                    : Ok   gcc (Alpine 5.3.0) 5.3.0
   2 alpine:3.5                    : Ok   gcc (Alpine 6.2.1) 6.2.1 20160822
   3 alpine:3.6                    : Ok   gcc (Alpine 6.3.0) 6.3.0
   4 alpine:3.7                    : Ok   gcc (Alpine 6.4.0) 6.4.0
   5 alpine:3.8                    : Ok   gcc (Alpine 6.4.0) 6.4.0
   6 alpine:3.9                    : Ok   gcc (Alpine 8.2.0) 8.2.0
   7 alpine:edge                   : Ok   gcc (Alpine 8.2.0) 8.2.0
   8 amazonlinux:1                 : Ok   gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-28)
   9 amazonlinux:2                 : Ok   gcc (GCC) 7.3.1 20180303 (Red Hat 7.3.1-5)
  10 android-ndk:r12b-arm          : Ok   arm-linux-androideabi-gcc (GCC) 4.9.x 20150123 (prerelease)
  11 android-ndk:r15c-arm          : Ok   arm-linux-androideabi-gcc (GCC) 4.9.x 20150123 (prerelease)
  12 centos:5                      : Ok   gcc (GCC) 4.1.2 20080704 (Red Hat 4.1.2-55)
  13 centos:6                      : Ok   gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-23)
  14 centos:7                      : Ok   gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-36)
  15 clearlinux:latest             : Ok   gcc (Clear Linux OS for Intel Architecture) 8.2.1 20180502
  16 debian:7                      : Ok   gcc (Debian 4.7.2-5) 4.7.2
  17 debian:8                      : Ok   gcc (Debian 4.9.2-10+deb8u2) 4.9.2
  18 debian:9                      : Ok   gcc (Debian 6.3.0-18+deb9u1) 6.3.0 20170516
  19 debian:experimental           : Ok   gcc (Debian 8.2.0-17) 8.2.1 20190204
  20 debian:experimental-x-arm64   : Ok   aarch64-linux-gnu-gcc (Debian 8.2.0-11) 8.2.0
  21 debian:experimental-x-mips    : Ok   mips-linux-gnu-gcc (Debian 8.2.0-11) 8.2.0
  22 debian:experimental-x-mips64  : Ok   mips64-linux-gnuabi64-gcc (Debian 8.2.0-16) 8.2.0
  23 fedora:20                     : Ok   gcc (GCC) 4.8.3 20140911 (Red Hat 4.8.3-7)
  24 fedora:21                     : Ok   gcc (GCC) 4.9.2 20150212 (Red Hat 4.9.2-6)
  25 fedora:22                     : Ok   gcc (GCC) 5.3.1 20160406 (Red Hat 5.3.1-6)
  26 fedora:23                     : Ok   gcc (GCC) 5.3.1 20160406 (Red Hat 5.3.1-6)
  27 fedora:24                     : Ok   gcc (GCC) 6.3.1 20161221 (Red Hat 6.3.1-1)
  28 fedora:24-x-ARC-uClibc        : Ok   arc-linux-gcc (ARCompact ISA Linux uClibc toolchain 2017.09-rc2) 7.1.1 20170710
  29 fedora:25                     : Ok   gcc (GCC) 6.4.1 20170727 (Red Hat 6.4.1-1)
  30 fedora:26                     : Ok   gcc (GCC) 7.3.1 20180130 (Red Hat 7.3.1-2)
  31 fedora:27                     : Ok   gcc (GCC) 7.3.1 20180712 (Red Hat 7.3.1-6)
  32 fedora:28                     : Ok   gcc (GCC) 8.2.1 20181215 (Red Hat 8.2.1-6)
  33 fedora:29                     : Ok   gcc (GCC) 8.2.1 20181215 (Red Hat 8.2.1-6)
  34 fedora:30                     : Ok   gcc (GCC) 9.0.1 20190203 (Red Hat 9.0.1-0.3)
  35 fedora:rawhide                : Ok   gcc (GCC) 9.0.0 20190119 (Red Hat 9.0.0-0.3)
  36 gentoo-stage3-amd64:latest    : Ok   gcc (Gentoo 7.3.0-r3 p1.4) 7.3.0
  37 mageia:5                      : Ok   gcc (GCC) 4.9.2
  38 mageia:6                      : Ok   gcc (Mageia 5.5.0-1.mga6) 5.5.0
  39 opensuse:13.2                 : Ok   gcc (SUSE Linux) 4.8.3 20140627 [gcc-4_8-branch revision 212064]
  40 opensuse:15.0                 : Ok   gcc (SUSE Linux) 7.3.1 20180323 [gcc-7-branch revision 258812]
  41 opensuse:15.1                 : Ok   gcc (SUSE Linux) 7.4.0
  42 opensuse:42.1                 : Ok   gcc (SUSE Linux) 4.8.5
  43 opensuse:42.2                 : Ok   gcc (SUSE Linux) 4.8.5
  44 opensuse:42.3                 : Ok   gcc (SUSE Linux) 4.8.5
  45 opensuse:tumbleweed           : Ok   gcc (SUSE Linux) 8.2.1 20190103 [gcc-8-branch revision 267549]
  46 oraclelinux:6                 : Ok   gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-23.0.1)
  47 oraclelinux:7                 : Ok   gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-36.0.1)
  48 ubuntu:12.04.5                : Ok   gcc (Ubuntu/Linaro 4.6.3-1ubuntu5) 4.6.3
  49 ubuntu:14.04.4                : Ok   gcc (Ubuntu 4.8.4-2ubuntu1~14.04.4) 4.8.4
  50 ubuntu:14.04.4-x-linaro-arm64 : Ok   aarch64-linux-gnu-gcc (Linaro GCC 5.5-2017.10) 5.5.0
  51 ubuntu:16.04                  : Ok   gcc (Ubuntu 5.4.0-6ubuntu1~16.04.11) 5.4.0 20160609
  52 ubuntu:16.04-x-arm            : Ok   arm-linux-gnueabihf-gcc (Ubuntu/Linaro 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609
  53 ubuntu:16.04-x-arm64          : Ok   aarch64-linux-gnu-gcc (Ubuntu/Linaro 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609
  54 ubuntu:16.04-x-powerpc        : Ok   powerpc-linux-gnu-gcc (Ubuntu 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609
  55 ubuntu:16.04-x-powerpc64      : Ok   powerpc64-linux-gnu-gcc (Ubuntu/IBM 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609
  56 ubuntu:16.04-x-powerpc64el    : Ok   powerpc64le-linux-gnu-gcc (Ubuntu/IBM 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609
  57 ubuntu:16.04-x-s390           : Ok   s390x-linux-gnu-gcc (Ubuntu 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609
  58 ubuntu:17.10                  : Ok   gcc (Ubuntu 7.2.0-8ubuntu3.2) 7.2.0
  59 ubuntu:18.04                  : Ok   gcc (Ubuntu 7.3.0-27ubuntu1~18.04) 7.3.0
  60 ubuntu:18.04-x-arm            : Ok   arm-linux-gnueabihf-gcc (Ubuntu/Linaro 7.3.0-27ubuntu1~18.04) 7.3.0
  61 ubuntu:18.04-x-arm64          : Ok   aarch64-linux-gnu-gcc (Ubuntu/Linaro 7.3.0-27ubuntu1~18.04) 7.3.0
  62 ubuntu:18.04-x-m68k           : Ok   m68k-linux-gnu-gcc (Ubuntu 7.3.0-27ubuntu1~18.04) 7.3.0
  63 ubuntu:18.04-x-powerpc        : Ok   powerpc-linux-gnu-gcc (Ubuntu 7.3.0-27ubuntu1~18.04) 7.3.0
  64 ubuntu:18.04-x-powerpc64      : Ok   powerpc64-linux-gnu-gcc (Ubuntu 7.3.0-27ubuntu1~18.04) 7.3.0
  65 ubuntu:18.04-x-powerpc64el    : Ok   powerpc64le-linux-gnu-gcc (Ubuntu 7.3.0-27ubuntu1~18.04) 7.3.0
  66 ubuntu:18.04-x-riscv64        : Ok   riscv64-linux-gnu-gcc (Ubuntu 7.3.0-27ubuntu1~18.04) 7.3.0
  67 ubuntu:18.04-x-s390           : Ok   s390x-linux-gnu-gcc (Ubuntu 7.3.0-27ubuntu1~18.04) 7.3.0
  68 ubuntu:18.04-x-sh4            : Ok   sh4-linux-gnu-gcc (Ubuntu 7.3.0-27ubuntu1~18.04) 7.3.0
  69 ubuntu:18.04-x-sparc64        : Ok   sparc64-linux-gnu-gcc (Ubuntu 7.3.0-27ubuntu1~18.04) 7.3.0
  70 ubuntu:18.10                  : Ok   gcc (Ubuntu 8.2.0-7ubuntu1) 8.2.0
  71 ubuntu:19.04                  : Ok   gcc (Ubuntu 8.2.0-20ubuntu1) 8.2.0
  72 ubuntu:19.04-x-alpha          : Ok   alpha-linux-gnu-gcc (Ubuntu 8.2.0-20ubuntu1) 8.2.0
  73 ubuntu:19.04-x-arm64          : Ok   aarch64-linux-gnu-gcc (Ubuntu/Linaro 8.2.0-20ubuntu1) 8.2.0
  74 ubuntu:19.04-x-hppa           : Ok   hppa-linux-gnu-gcc (Ubuntu 8.2.0-20ubuntu1) 8.2.0
  #
  # uname -a
  Linux quaco 4.20.6-200.fc29.x86_64 #1 SMP Thu Jan 31 15:50:43 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux
  # git log --oneline -1
  b4409ae112ca perf tools: Make rm_rf() remove single file
  # perf version --build-options
  perf version 5.0.rc5.gb4409a
                   dwarf: [ on  ]  # HAVE_DWARF_SUPPORT
      dwarf_getlocations: [ on  ]  # HAVE_DWARF_GETLOCATIONS_SUPPORT
                   glibc: [ on  ]  # HAVE_GLIBC_SUPPORT
                    gtk2: [ on  ]  # HAVE_GTK2_SUPPORT
           syscall_table: [ on  ]  # HAVE_SYSCALL_TABLE_SUPPORT
                  libbfd: [ on  ]  # HAVE_LIBBFD_SUPPORT
                  libelf: [ on  ]  # HAVE_LIBELF_SUPPORT
                 libnuma: [ on  ]  # HAVE_LIBNUMA_SUPPORT
  numa_num_possible_cpus: [ on  ]  # HAVE_LIBNUMA_SUPPORT
                 libperl: [ on  ]  # HAVE_LIBPERL_SUPPORT
               libpython: [ on  ]  # HAVE_LIBPYTHON_SUPPORT
                libslang: [ on  ]  # HAVE_SLANG_SUPPORT
               libcrypto: [ on  ]  # HAVE_LIBCRYPTO_SUPPORT
               libunwind: [ on  ]  # HAVE_LIBUNWIND_SUPPORT
      libdw-dwarf-unwind: [ on  ]  # HAVE_DWARF_SUPPORT
                    zlib: [ on  ]  # HAVE_ZLIB_SUPPORT
                    lzma: [ on  ]  # HAVE_LZMA_SUPPORT
               get_cpuid: [ on  ]  # HAVE_AUXTRACE_SUPPORT
                     bpf: [ on  ]  # HAVE_LIBBPF_SUPPORT
  # perf test
   1: vmlinux symtab matches kallsyms                       : Ok
   2: Detect openat syscall event                           : Ok
   3: Detect openat syscall event on all cpus               : Ok
   4: Read samples using the mmap interface                 : Ok
   5: Test data source output                               : Ok
   6: Parse event definition strings                        : Ok
   7: Simple expression parser                              : Ok
   8: PERF_RECORD_* events & perf_sample fields             : Ok
   9: Parse perf pmu format                                 : Ok
  10: DSO data read                                         : Ok
  11: DSO data cache                                        : Ok
  12: DSO data reopen                                       : Ok
  13: Roundtrip evsel->name                                 : Ok
  14: Parse sched tracepoints fields                        : Ok
  15: syscalls:sys_enter_openat event fields                : Ok
  16: Setup struct perf_event_attr                          : Ok
  17: Match and link multiple hists                         : Ok
  18: 'import perf' in python                               : Ok
  19: Breakpoint overflow signal handler                    : Ok
  20: Breakpoint overflow sampling                          : Ok
  21: Breakpoint accounting                                 : Ok
  22: Watchpoint                                            :
  22.1: Read Only Watchpoint                                : Skip
  22.2: Write Only Watchpoint                               : Ok
  22.3: Read / Write Watchpoint                             : Ok
  22.4: Modify Watchpoint                                   : Ok
  23: Number of exit events of a simple workload            : Ok
  24: Software clock events period values                   : Ok
  25: Object code reading                                   : Ok
  26: Sample parsing                                        : Ok
  27: Use a dummy software event to keep tracking           : Ok
  28: Parse with no sample_id_all bit set                   : Ok
  29: Filter hist entries                                   : Ok
  30: Lookup mmap thread                                    : Ok
  31: Share thread mg                                       : Ok
  32: Sort output of hist entries                           : Ok
  33: Cumulate child hist entries                           : Ok
  34: Track with sched_switch                               : Ok
  35: Filter fds with revents mask in a fdarray             : Ok
  36: Add fd to a fdarray, making it autogrow               : Ok
  37: kmod_path__parse                                      : Ok
  38: Thread map                                            : Ok
  39: LLVM search and compile                               :
  39.1: Basic BPF llvm compile                              : Ok
  39.2: kbuild searching                                    : Ok
  39.3: Compile source for BPF prologue generation          : Ok
  39.4: Compile source for BPF relocation                   : Ok
  40: Session topology                                      : Ok
  41: BPF filter                                            :
  41.1: Basic BPF filtering                                 : Ok
  41.2: BPF pinning                                         : Ok
  41.3: BPF prologue generation                             : Ok
  41.4: BPF relocation checker                              : Ok
  42: Synthesize thread map                                 : Ok
  43: Remove thread map                                     : Ok
  44: Synthesize cpu map                                    : Ok
  45: Synthesize stat config                                : Ok
  46: Synthesize stat                                       : Ok
  47: Synthesize stat round                                 : Ok
  48: Synthesize attr update                                : Ok
  49: Event times                                           : Ok
  50: Read backward ring buffer                             : Ok
  51: Print cpu map                                         : Ok
  52: Probe SDT events                                      : Ok
  53: is_printable_array                                    : Ok
  54: Print bitmap                                          : Ok
  55: perf hooks                                            : Ok
  56: builtin clang support                                 : Skip (not compiled in)
  57: unit_number__scnprintf                                : Ok
  58: mem2node                                              : Ok
  59: x86 rdpmc                                             : Ok
  60: Convert perf time to TSC                              : Ok
  61: DWARF unwind                                          : Ok
  62: x86 instruction decoder - new instructions            : Ok
  63: x86 bp modify                                         : Ok
  64: probe libc's inet_pton & backtrace it with ping       : Ok
  65: Use vfs_getname probe to get syscall args filenames   : Ok
  66: Add vfs_getname probe to get syscall args filenames   : Ok
  67: Check open filename arg using perf trace + vfs_getname: Ok
  $ make -C tools/perf build-test
  make: Entering directory '/home/acme/git/perf/tools/perf'
  - tarpkg: ./tests/perf-targz-src-pkg .
              make_no_libelf_O: make NO_LIBELF=1
       make_util_pmu_bison_o_O: make util/pmu-bison.o
                 make_perf_o_O: make perf.o
            make_install_bin_O: make install-bin
           make_no_libbionic_O: make NO_LIBBIONIC=1
                   make_help_O: make help
            make_no_demangle_O: make NO_DEMANGLE=1
             make_no_scripts_O: make NO_LIBPYTHON=1 NO_LIBPERL=1
         make_with_clangllvm_O: make LIBCLANGLLVM=1
                make_install_O: make install
           make_no_backtrace_O: make NO_BACKTRACE=1
           make_no_libunwind_O: make NO_LIBUNWIND=1
             make_no_libperl_O: make NO_LIBPERL=1
            make_no_libaudit_O: make NO_LIBAUDIT=1
                make_minimal_O: make NO_LIBPERL=1 NO_LIBPYTHON=1 NO_NEWT=1 NO_GTK2=1 NO_DEMANGLE=1 NO_LIBELF=1 NO_LIBUNWIND=1 NO_BACKTRACE=1 NO_LIBNUMA=1 NO_LIBAUDIT=1 NO_LIBBIONIC=1 NO_LIBDW_DWARF_UNWIND=1 NO_AUXTRACE=1 NO_LIBBPF=1 NO_LIBCRYPTO=1 NO_SDT=1 NO_JVMTI=1
               make_no_slang_O: make NO_SLANG=1
              make_no_libbpf_O: make NO_LIBBPF=1
                 make_cscope_O: make cscope
                make_no_newt_O: make NO_NEWT=1
           make_no_libpython_O: make NO_LIBPYTHON=1
                  make_debug_O: make DEBUG=1
             make_util_map_o_O: make util/map.o
                   make_tags_O: make tags
  make_no_libdw_dwarf_unwind_O: make NO_LIBDW_DWARF_UNWIND=1
              make_clean_all_O: make clean all
                  make_no_ui_O: make NO_NEWT=1 NO_SLANG=1 NO_GTK2=1
             make_no_libnuma_O: make NO_LIBNUMA=1
                 make_static_O: make LDFLAGS=-static
   make_install_prefix_slash_O: make install prefix=/tmp/krava/
                make_no_gtk2_O: make NO_GTK2=1
        make_with_babeltrace_O: make LIBBABELTRACE=1
         make_install_prefix_O: make install prefix=/tmp/krava
            make_no_auxtrace_O: make NO_AUXTRACE=1
                    make_doc_O: make doc
                   make_pure_O: make
  OK
  make: Leaving directory '/home/acme/git/perf/tools/perf'
  $ 
author maintainer
#1 lore raw

PATCH 1/17 perf tests shell: Skip trace+probe_vfs_getname.sh if built without trace support

From: Tommi Rantala <[email protected]>

If perf was built without trace support, the trace+probe_vfs_getname.sh 'perf test' entry fails:

  # perf trace -h
  perf: 'trace' is not a perf-command. See 'perf --help'
  # perf test 64
  64: Check open filename arg using perf trace + vfs_getname: FAILED!

Check trace support, so that we'll skip the test in that case:

  # perf test 64
  64: Check open filename arg using perf trace + vfs_getname: Skip
Cc Alexander Shishkin <[email protected]>
Cc Hendrik Brueckner <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Kim Phillips <[email protected]>
Cc Michael Petlan <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Peter Zijlstra <[email protected]>
---
 tools/perf/tests/shell/lib/probe.sh               | 5 +++++
 tools/perf/tests/shell/trace+probe_vfs_getname.sh | 1 +
 2 files changed, 6 insertions(+)
tools/perf/tests/shell/lib/probe.sh +5 -0
index 6293cc660947..e37787be672b 100644
@@ -4,3 +4,8 @@ skip_if_no_perf_probe() {
 	perf probe 2>&1 | grep -q 'is not a perf-command' && return 2
 	return 0
 }
+
+skip_if_no_perf_trace() {
+	perf trace -h 2>&1 | grep -q -e 'is not a perf-command' -e 'trace command not available' && return 2
+	return 0
+}
tools/perf/tests/shell/trace+probe_vfs_getname.sh +1 -0
index 50109f27ca07..147efeb6b195 100755
@@ -12,6 +12,7 @@
 . $(dirname $0)/lib/probe.sh
 
 skip_if_no_perf_probe || exit 2
+skip_if_no_perf_trace || exit 2
 
 . $(dirname $0)/lib/probe_vfs_getname.sh
 
Signature · 1 line
2.19.1
author maintainer
#2 lore raw

PATCH 2/17 perf header: Fix wrong node write in NUMA_TOPOLOGY feature

In reply to Arnaldo Carvalho de Melo

From: Jiri Olsa <[email protected]>

We are currently passing the node index instead of the real node number.

Cc Alexander Shishkin <[email protected]>
Cc Peter Zijlstra <[email protected]>
Fixes fbe96f29ce4b ("perf tools: Make perf.data more self-descriptive (v8)"
---
 tools/perf/util/header.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/util/header.c +1 -1
index 61ce197c5362..c66f26ec557a 100644
@@ -879,7 +879,7 @@ static int write_numa_topology(struct feat_fd *ff,
 		if (ret < 0)
 			break;
 
-		ret = write_topo_node(ff, i);
+		ret = write_topo_node(ff, j);
 		if (ret < 0)
 			break;
 	}
Signature · 1 line
2.19.1
author maintainer
#3 lore raw

PATCH 3/17 perf tools: Add cpu_topology object

In reply to Arnaldo Carvalho de Melo

From: Jiri Olsa <[email protected]>

Make struct cpu_topo global and rename it to 'struct cpu_topology', so that it can be used from the 'perf record' command in the following patches.

Add the following interface functions to load/free cpu topology details:

  struct cpu_topology *cpu_topology__new(void);
  void cpu_topology__delete(struct cpu_topology *tp);

Move it to a separate source file cputopo.c together with numa related object in the following patches.

No functional change, the new interface will be used in upcoming changes.

---
 tools/perf/util/Build     |   1 +
 tools/perf/util/cputopo.c | 144 ++++++++++++++++++++++++++++++++++++
 tools/perf/util/cputopo.h |  17 +++++
 tools/perf/util/header.c  | 150 +-------------------------------------
 4 files changed, 166 insertions(+), 146 deletions(-)
 create mode 100644 tools/perf/util/cputopo.c
 create mode 100644 tools/perf/util/cputopo.h
tools/perf/util/Build +1 -0
index ca0741c91903..3008d49fa587 100644
@@ -69,6 +69,7 @@ perf-y += hist.o
 perf-y += util.o
 perf-y += xyarray.o
 perf-y += cpumap.o
+perf-y += cputopo.o
 perf-y += cgroup.o
 perf-y += target.o
 perf-y += rblist.o
tools/perf/util/cputopo.c new file +144 -0

144 lines not on this page · raw message

tools/perf/util/cputopo.h new file +17 -0
new file mode 100644
index 000000000000..4b5f4112b6f8
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __PERF_CPUTOPO_H
+#define __PERF_CPUTOPO_H
+
+#include <linux/types.h>
+
+struct cpu_topology {
+	u32	  core_sib;
+	u32	  thread_sib;
+	char	**core_siblings;
+	char	**thread_siblings;
+};
+
+struct cpu_topology *cpu_topology__new(void);
+void cpu_topology__delete(struct cpu_topology *tp);
+
+#endif /* __PERF_CPUTOPO_H */
tools/perf/util/header.c +4 -146

175 lines not on this page · raw message

Signature · 1 line
2.19.1
author maintainer
#4 lore raw

PATCH 4/17 perf tools: Add numa_topology object

In reply to Arnaldo Carvalho de Melo

From: Jiri Olsa <[email protected]>

Add the numa_topology object to return the list of numa nodes together with their cpus. It will replace the numa code in header.c and will be used from 'perf record' in the following patches.

Add the following interface functions to load numa details:

  struct numa_topology *numa_topology__new(void);
  void numa_topology__delete(struct numa_topology *tp);

And replace the current (copied) local interface, with no functional changes.

---
 tools/perf/util/cputopo.c | 118 +++++++++++++++++++++++++++++++++++++
 tools/perf/util/cputopo.h |  16 +++++
 tools/perf/util/header.c  | 119 +++++++++-----------------------------
 3 files changed, 160 insertions(+), 93 deletions(-)
tools/perf/util/cputopo.c +118 -0
index 84470ed4e707..83ffca2ea9ee 100644
@@ -1,9 +1,11 @@
 // SPDX-License-Identifier: GPL-2.0
 #include <sys/param.h>
+#include <inttypes.h>
 
 #include "cputopo.h"
 #include "cpumap.h"
 #include "util.h"
+#include "env.h"
 
 
 #define CORE_SIB_FMT \
@@ -142,3 +144,119 @@ struct cpu_topology *cpu_topology__new(void)
 	}
 	return tp;
 }
+
+static int load_numa_node(struct numa_topology_node *node, int nr)
+{
+	char str[MAXPATHLEN];
+	char field[32];
+	char *buf = NULL, *p;
+	size_t len = 0;
+	int ret = -1;
+	FILE *fp;
+	u64 mem;
+
+	node->node = (u32) nr;
+
+	sprintf(str, "/sys/devices/system/node/node%d/meminfo", nr);
+	fp = fopen(str, "r");
+	if (!fp)
+		return -1;
+
+	while (getline(&buf, &len, fp) > 0) {
+		/* skip over invalid lines */
+		if (!strchr(buf, ':'))
+			continue;
+		if (sscanf(buf, "%*s %*d %31s %"PRIu64, field, &mem) != 2)
+			goto err;
+		if (!strcmp(field, "MemTotal:"))
+			node->mem_total = mem;
+		if (!strcmp(field, "MemFree:"))
+			node->mem_free = mem;
+		if (node->mem_total && node->mem_free)
+			break;
+	}
+
+	fclose(fp);
+	fp = NULL;
+
+	sprintf(str, "/sys/devices/system/node/node%d/cpulist", nr);
+
+	fp = fopen(str, "r");
+	if (!fp)
+		return -1;
+
+	if (getline(&buf, &len, fp) <= 0)
+		goto err;
+
+	p = strchr(buf, '\n');
+	if (p)
+		*p = '\0';
+
+	node->cpus = buf;
+	fclose(fp);
+	return 0;
+
+err:
+	free(buf);
+	if (fp)
+		fclose(fp);
+	return ret;
+}
+
+struct numa_topology *numa_topology__new(void)
+{
+	struct cpu_map *node_map = NULL;
+	struct numa_topology *tp = NULL;
+	char *buf = NULL;
+	size_t len = 0;
+	u32 nr, i;
+	FILE *fp;
+	char *c;
+
+	fp = fopen("/sys/devices/system/node/online", "r");
+	if (!fp)
+		return NULL;
+
+	if (getline(&buf, &len, fp) <= 0)
+		goto out;
+
+	c = strchr(buf, '\n');
+	if (c)
+		*c = '\0';
+
+	node_map = cpu_map__new(buf);
+	if (!node_map)
+		goto out;
+
+	nr = (u32) node_map->nr;
+
+	tp = zalloc(sizeof(*tp) + sizeof(tp->nodes[0])*nr);
+	if (!tp)
+		goto out;
+
+	tp->nr = nr;
+
+	for (i = 0; i < nr; i++) {
+		if (load_numa_node(&tp->nodes[i], node_map->map[i])) {
+			numa_topology__delete(tp);
+			tp = NULL;
+			break;
+		}
+	}
+
+out:
+	free(buf);
+	fclose(fp);
+	cpu_map__put(node_map);
+	return tp;
+}
+
+void numa_topology__delete(struct numa_topology *tp)
+{
+	u32 i;
+
+	for (i = 0; i < tp->nr; i++)
+		free(tp->nodes[i].cpus);
+
+	free(tp);
+}
tools/perf/util/cputopo.h +16 -0
index 4b5f4112b6f8..47a97e71acdf 100644
@@ -3,6 +3,7 @@
 #define __PERF_CPUTOPO_H
 
 #include <linux/types.h>
+#include "env.h"
 
 struct cpu_topology {
 	u32	  core_sib;
@@ -11,7 +12,22 @@ struct cpu_topology {
 	char	**thread_siblings;
 };
 
+struct numa_topology_node {
+	char		*cpus;
+	u32		 node;
+	u64		 mem_total;
+	u64		 mem_free;
+};
+
+struct numa_topology {
+	u32				nr;
+	struct numa_topology_node	nodes[0];
+};
+
 struct cpu_topology *cpu_topology__new(void);
 void cpu_topology__delete(struct cpu_topology *tp);
 
+struct numa_topology *numa_topology__new(void);
+void numa_topology__delete(struct numa_topology *tp);
+
 #endif /* __PERF_CPUTOPO_H */
tools/perf/util/header.c +26 -93

138 lines not on this page · raw message

Signature · 1 line
2.19.1
author maintainer
#5 lore raw

PATCH 5/17 perf tools: Use sysfs__mountpoint() when reading cpu topology

In reply to Arnaldo Carvalho de Melo

From: Jiri Olsa <[email protected]>

Use sysfs__mountpoint() when reading sysfs files to obtain cpu/numa topologies.

Also use scnprintf instead of sprintf as suggested by Namhyung.

---
 tools/perf/util/cputopo.c | 29 ++++++++++++++++++++++-------
 1 file changed, 22 insertions(+), 7 deletions(-)
tools/perf/util/cputopo.c +22 -7
index 83ffca2ea9ee..ece0710249d4 100644
@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 #include <sys/param.h>
 #include <inttypes.h>
+#include <api/fs/fs.h>
 
 #include "cputopo.h"
 #include "cpumap.h"
@@ -9,9 +10,15 @@
 
 
 #define CORE_SIB_FMT \
-	"/sys/devices/system/cpu/cpu%d/topology/core_siblings_list"
+	"%s/devices/system/cpu/cpu%d/topology/core_siblings_list"
 #define THRD_SIB_FMT \
-	"/sys/devices/system/cpu/cpu%d/topology/thread_siblings_list"
+	"%s/devices/system/cpu/cpu%d/topology/thread_siblings_list"
+#define NODE_ONLINE_FMT \
+	"%s/devices/system/node/online"
+#define NODE_MEMINFO_FMT \
+	"%s/devices/system/node/node%d/meminfo"
+#define NODE_CPULIST_FMT \
+	"%s/devices/system/node/node%d/cpulist"
 
 static int build_cpu_topology(struct cpu_topology *tp, int cpu)
 {
@@ -23,7 +30,8 @@ static int build_cpu_topology(struct cpu_topology *tp, int cpu)
 	u32 i = 0;
 	int ret = -1;
 
-	sprintf(filename, CORE_SIB_FMT, cpu);
+	scnprintf(filename, MAXPATHLEN, CORE_SIB_FMT,
+		  sysfs__mountpoint(), cpu);
 	fp = fopen(filename, "r");
 	if (!fp)
 		goto try_threads;
@@ -50,7 +58,8 @@ static int build_cpu_topology(struct cpu_topology *tp, int cpu)
 	ret = 0;
 
 try_threads:
-	sprintf(filename, THRD_SIB_FMT, cpu);
+	scnprintf(filename, MAXPATHLEN, THRD_SIB_FMT,
+		  sysfs__mountpoint(), cpu);
 	fp = fopen(filename, "r");
 	if (!fp)
 		goto done;
@@ -157,7 +166,8 @@ static int load_numa_node(struct numa_topology_node *node, int nr)
 
 	node->node = (u32) nr;
 
-	sprintf(str, "/sys/devices/system/node/node%d/meminfo", nr);
+	scnprintf(str, MAXPATHLEN, NODE_MEMINFO_FMT,
+		  sysfs__mountpoint(), nr);
 	fp = fopen(str, "r");
 	if (!fp)
 		return -1;
@@ -179,7 +189,8 @@ static int load_numa_node(struct numa_topology_node *node, int nr)
 	fclose(fp);
 	fp = NULL;
 
-	sprintf(str, "/sys/devices/system/node/node%d/cpulist", nr);
+	scnprintf(str, MAXPATHLEN, NODE_CPULIST_FMT,
+		  sysfs__mountpoint(), nr);
 
 	fp = fopen(str, "r");
 	if (!fp)
@@ -207,13 +218,17 @@ struct numa_topology *numa_topology__new(void)
 {
 	struct cpu_map *node_map = NULL;
 	struct numa_topology *tp = NULL;
+	char path[MAXPATHLEN];
 	char *buf = NULL;
 	size_t len = 0;
 	u32 nr, i;
 	FILE *fp;
 	char *c;
 
-	fp = fopen("/sys/devices/system/node/online", "r");
+	scnprintf(path, MAXPATHLEN, NODE_ONLINE_FMT,
+		  sysfs__mountpoint());
+
+	fp = fopen(path, "r");
 	if (!fp)
 		return NULL;
 
Signature · 1 line
2.19.1
author maintainer
#6 lore raw

PATCH 6/17 perf report: Don't shadow inlined symbol with different addr range

In reply to Arnaldo Carvalho de Melo

From: He Kuang <[email protected]>

We can't assume inlined symbols with the same name are equal, because their address range may be different. This will cause the symbols with different addresses be shadowed when adding to the hist entry, and lead to ERANGE error when checking the symbol address during sample parse, the addr should be within the range of [sym.start, sym.end].

The error message is like: "0x36aea60 [0x8]: failed to process type: 68".

The second parameter of symbol__new() is the length of the fake symbol for the inline frame, which is the subtraction of the end and start address of base_sym.

Cc Alexander Shishkin <[email protected]>
Cc Milian Wolff <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Peter Zijlstra <[email protected]>
Fixes aa441895f7b4 ("perf report: Compare symbol name for inlined frames when sorting")
---
 tools/perf/util/sort.c    | 10 ++++++++--
 tools/perf/util/srcline.c |  2 +-
 2 files changed, 9 insertions(+), 3 deletions(-)
tools/perf/util/sort.c +8 -2
index 2b6c1ccb878c..d2299e912e59 100644
@@ -231,8 +231,14 @@ static int64_t _sort__sym_cmp(struct symbol *sym_l, struct symbol *sym_r)
 	if (sym_l == sym_r)
 		return 0;
 
-	if (sym_l->inlined || sym_r->inlined)
-		return strcmp(sym_l->name, sym_r->name);
+	if (sym_l->inlined || sym_r->inlined) {
+		int ret = strcmp(sym_l->name, sym_r->name);
+
+		if (ret)
+			return ret;
+		if ((sym_l->start <= sym_r->end) && (sym_l->end >= sym_r->start))
+			return 0;
+	}
 
 	if (sym_l->start != sym_r->start)
 		return (int64_t)(sym_r->start - sym_l->start);
tools/perf/util/srcline.c +1 -1
index 00f215580b5a..10ca1533937e 100644
@@ -104,7 +104,7 @@ static struct symbol *new_inline_sym(struct dso *dso,
 	} else {
 		/* create a fake symbol for the inline frame */
 		inline_sym = symbol__new(base_sym ? base_sym->start : 0,
-					 base_sym ? base_sym->end : 0,
+					 base_sym ? (base_sym->end - base_sym->start) : 0,
 					 base_sym ? base_sym->binding : 0,
 					 base_sym ? base_sym->type : 0,
 					 funcname);
Signature · 1 line
2.19.1
author maintainer
#7 lore raw

PATCH 7/17 perf doc: Fix HEADER_CMDLINE description in perf.data documentation

In reply to Arnaldo Carvalho de Melo

From: Jonas Rabenstein <[email protected]>

The content of the HEADER_CMDLINE feature header is a perf_header_string_list of the argument vector and not a perf_header_string of the commandline.

---
 tools/perf/Documentation/perf.data-file-format.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/Documentation/perf.data-file-format.txt +1 -1
index dfb218feaad9..5f9a3924830b 100644
@@ -131,7 +131,7 @@ An uint64_t with the total memory in bytes.
 
 	HEADER_CMDLINE = 11,
 
-A perf_header_string with the perf command line used to collect the data.
+A perf_header_string_list with the perf arg-vector used to collect the data.
 
 	HEADER_EVENT_DESC = 12,
 
Signature · 1 line
2.19.1
author maintainer
#8 lore raw

PATCH 8/17 perf doc: Fix documentation of the Flags section in perf.data

In reply to Arnaldo Carvalho de Melo

From: Jonas Rabenstein <[email protected]>

According to the current documentation the flags section is placed after the file header itself but the code assumes to find the flags section after the data section. This change updates the documentation to that assumption.

---
 tools/perf/Documentation/perf.data-file-format.txt | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)
tools/perf/Documentation/perf.data-file-format.txt +4 -5
index 5f9a3924830b..593ef49b273c 100644
@@ -43,11 +43,10 @@ struct perf_file_section {
 
 Flags section:
 
-The header is followed by different optional headers, described by the bits set
-in flags. Only headers for which the bit is set are included. Each header
-consists of a perf_file_section located after the initial header.
-The respective perf_file_section points to the data of the additional
-header and defines its size.
+For each of the optional features a perf_file_section it placed after the data
+section if the feature bit is set in the perf_header flags bitset. The
+respective perf_file_section points to the data of the additional header and
+defines its size.
 
 Some headers consist of strings, which are defined like this:
 
Signature · 1 line
2.19.1
author maintainer
#9 lore raw

PATCH 9/17 perf test: Fix failure of 'evsel-tp-sched' test on s390

In reply to Arnaldo Carvalho de Melo

From: Thomas Richter <[email protected]>

Commit 489338a717a0 ("perf tests evsel-tp-sched: Fix bitwise operator") causes test case 14 "Parse sched tracepoints fields" to fail on s390.

This test succeeds on x86.

In fact this test now fails on all architectures with type char treated as type unsigned char.

The root cause is the signed-ness of character arrays in the tracepoints sched_switch for structure members prev_comm and next_comm.

On s390 the output of:

[root@m35lp76 perf]# cat /sys/kernel/debug/tracing/events/sched/sched_switch/format
name: sched_switch
ID: 287
format:
field:unsigned short common_type; offset:0; size:2; signed:0;
...
field:char prev_comm[16]; offset:8; size:16; signed:0;
...
field:char next_comm[16]; offset:40; size:16; signed:0;

reveals the character arrays prev_comm and next_comm are per default unsigned char and have values in the range of 0..255.

On x86 both fields are signed as this output shows:

 [root@f29]# cat /sys/kernel/debug/tracing/events/sched/sched_switch/format
 name: sched_switch
 ID: 287
 format:
   field:unsigned short common_type; offset:0; size:2;	signed:0;
   ...
   field:char prev_comm[16]; offset:8; size:16;	signed:1;
   ...
   field:char next_comm[16]; offset:40; size:16; signed:1;

and the character arrays prev_comm and next_comm are per default signed char and have values in the range of -1..127. The implementation of type char is architecture specific.

Since the character arrays in both tracepoints sched_switch and sched_wakeup should contain ascii characters, simply omit the check for signedness in the test case.

Output before:

[root@m35lp76 perf]# ./perf test -F 14
14: Parse sched tracepoints fields :
--- start ---
sched:sched_switch: "prev_comm" signedness(0) is wrong, should be 1 sched:sched_switch: "next_comm" signedness(0) is wrong, should be 1 sched:sched_wakeup: "comm" signedness(0) is wrong, should be 1 ---- end ----
14: Parse sched tracepoints fields : FAILED!
[root@m35lp76 perf]#

Output after:

[root@m35lp76 perf]# ./perf test -Fv 14
14: Parse sched tracepoints fields :
--- start ---
---- end ----
Parse sched tracepoints fields: Ok
[root@m35lp76 perf]#

Fixes 489338a717a0 ("perf tests evsel-tp-sched: Fix bitwise operator")
---
 tools/perf/tests/evsel-tp-sched.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
tools/perf/tests/evsel-tp-sched.c +3 -3
index 5cbba70bcdd0..ea7acf403727 100644
@@ -43,7 +43,7 @@ int test__perf_evsel__tp_sched_test(struct test *test __maybe_unused, int subtes
 		return -1;
 	}
 
-	if (perf_evsel__test_field(evsel, "prev_comm", 16, true))
+	if (perf_evsel__test_field(evsel, "prev_comm", 16, false))
 		ret = -1;
 
 	if (perf_evsel__test_field(evsel, "prev_pid", 4, true))
@@ -55,7 +55,7 @@ int test__perf_evsel__tp_sched_test(struct test *test __maybe_unused, int subtes
 	if (perf_evsel__test_field(evsel, "prev_state", sizeof(long), true))
 		ret = -1;
 
-	if (perf_evsel__test_field(evsel, "next_comm", 16, true))
+	if (perf_evsel__test_field(evsel, "next_comm", 16, false))
 		ret = -1;
 
 	if (perf_evsel__test_field(evsel, "next_pid", 4, true))
@@ -73,7 +73,7 @@ int test__perf_evsel__tp_sched_test(struct test *test __maybe_unused, int subtes
 		return -1;
 	}
 
-	if (perf_evsel__test_field(evsel, "comm", 16, true))
+	if (perf_evsel__test_field(evsel, "comm", 16, false))
 		ret = -1;
 
 	if (perf_evsel__test_field(evsel, "pid", 4, true))
Signature · 1 line
2.19.1
author maintainer
#10 lore raw

PATCH 10/17 perf bpf: Add bpf_map dumper

In reply to Arnaldo Carvalho de Melo

From: Arnaldo Carvalho de Melo <[email protected]>

At some point I'll suggest moving this to libbpf, for now I'll experiment with ways to dump BPF maps set by events in 'perf trace', starting with a very basic dumper for the current very limited needs of the augmented_raw_syscalls code: dumping booleans.

Having functions that apply to the map keys and values and do table lookup in things like syscall id to string tables should come next.

Cc Adrian Hunter <[email protected]>
Cc Alexei Starovoitov <[email protected]>
Cc Daniel Borkmann <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Martin KaFai Lau <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Yonghong Song <[email protected]>
---
 tools/perf/util/Build     |  1 +
 tools/perf/util/bpf_map.c | 72 +++++++++++++++++++++++++++++++++++++++
 tools/perf/util/bpf_map.h | 22 ++++++++++++
 3 files changed, 95 insertions(+)
 create mode 100644 tools/perf/util/bpf_map.c
 create mode 100644 tools/perf/util/bpf_map.h
tools/perf/util/Build +1 -0
index 3008d49fa587..8dd3102301ea 100644
@@ -115,6 +115,7 @@ perf-y += branch.o
 perf-y += mem2node.o
 
 perf-$(CONFIG_LIBBPF) += bpf-loader.o
+perf-$(CONFIG_LIBBPF) += bpf_map.o
 perf-$(CONFIG_BPF_PROLOGUE) += bpf-prologue.o
 perf-$(CONFIG_LIBELF) += symbol-elf.o
 perf-$(CONFIG_LIBELF) += probe-file.o
tools/perf/util/bpf_map.c new file +72 -0

72 lines not on this page · raw message

tools/perf/util/bpf_map.h new file +22 -0
new file mode 100644
index 000000000000..d6abd5e47af8
@@ -0,0 +1,22 @@
+// SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause)
+#ifndef __PERF_BPF_MAP_H
+#define __PERF_BPF_MAP_H 1
+
+#include <stdio.h>
+#include <linux/compiler.h>
+struct bpf_map;
+
+#ifdef HAVE_LIBBPF_SUPPORT
+
+int bpf_map__fprintf(struct bpf_map *map, FILE *fp);
+
+#else
+
+static inline int bpf_map__fprintf(struct bpf_map *map __maybe_unused, FILE *fp __maybe_unused)
+{
+	return 0;
+}
+
+#endif // HAVE_LIBBPF_SUPPORT
+
+#endif // __PERF_BPF_MAP_H
Signature · 1 line
2.19.1
author maintainer
#11 lore raw

PATCH 11/17 perf trace: Allow dumping a BPF map after setting up BPF events

In reply to Arnaldo Carvalho de Melo

From: Arnaldo Carvalho de Melo <[email protected]>

Initial use case:

Dumping the maps setup by tools/perf/examples/bpf/augmented_raw_syscalls.c, which so far are just booleans, showing just non-zeroed entries:

  # cat ~/.perfconfig
  [llvm]
	dump-obj = true
	clang-opt = -g
  [trace]
	#add_events = /home/acme/git/perf/tools/perf/examples/bpf/augmented_raw_syscalls.o
	add_events = /wb/augmented_raw_syscalls.o
  $ date
  Tue Feb 19 16:29:33 -03 2019
  $ ls -la /wb/augmented_raw_syscalls.o
  -rwxr-xr-x. 1 root root 14048 Jan 24 12:09 /wb/augmented_raw_syscalls.o
  $ file /wb/augmented_raw_syscalls.o
  /wb/augmented_raw_syscalls.o: ELF 64-bit LSB relocatable, eBPF, version 1 (SYSV), with debug_info, not stripped
  $
  # trace -e recvmmsg,sendmmsg --map-dump foobar
  ERROR: BPF map "foobar" not found
  # trace -e recvmmsg,sendmmsg --map-dump filtered_pids
  ERROR: BPF map "filtered_pids" not found
  # trace -e recvmmsg,sendmmsg --map-dump pids_filtered
  [2583] = 1,
  [2267] = 1,
  ^Z
  [1]+  Stopped                 trace -e recvmmsg,sendmmsg --map-dump pids_filtered
  # pidof trace
  2267
  # ps ax|grep gnome-terminal|grep -v grep
  2583 ?        Ssl   58:33 /usr/libexec/gnome-terminal-server
  ^C
  # trace -e recvmmsg,sendmmsg --map-dump syscalls
  [299] = 1,
  [307] = 1,
  ^C
  # grep x64_recvmmsg arch/x86/entry/syscalls/syscall_64.tbl
  299	64	recvmmsg		__x64_sys_recvmmsg
  # grep x64_sendmmsg arch/x86/entry/syscalls/syscall_64.tbl
  307	64	sendmmsg		__x64_sys_sendmmsg
  #

Next step probably will be something like 'perf stat's --interval-print and --interval-clear.

Cc Adrian Hunter <[email protected]>
Cc Alexei Starovoitov <[email protected]>
Cc Daniel Borkmann <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Martin KaFai Lau <[email protected]>
Cc Namhyung Kim <[email protected]>
Cc Yonghong Song <[email protected]>
---
 tools/perf/Documentation/perf-trace.txt |  8 ++++++++
 tools/perf/builtin-trace.c              | 19 +++++++++++++++++++
 2 files changed, 27 insertions(+)
tools/perf/Documentation/perf-trace.txt +8 -0
index 631e687be4eb..fc6e43262c41 100644
@@ -210,6 +210,14 @@ the thread executes on the designated CPUs. Default is to monitor all CPUs.
 	may happen, for instance, when a thread gets migrated to a different CPU
 	while processing a syscall.
 
+--map-dump::
+	Dump BPF maps setup by events passed via -e, for instance the augmented_raw_syscalls
+	living in tools/perf/examples/bpf/augmented_raw_syscalls.c. For now this
+	dumps just boolean map values and integer keys, in time this will print in hex
+	by default and use BTF when available, as well as use functions to do pretty
+	printing using the existing 'perf trace' syscall arg beautifiers to map integer
+	arguments to strings (pid to comm, syscall id to syscall name, etc).
+
 
 PAGEFAULTS
 ----------
tools/perf/builtin-trace.c +19 -0
index 68a01e624ad3..1a11fe656afc 100644
@@ -19,6 +19,7 @@
 #include <traceevent/event-parse.h>
 #include <api/fs/tracing_path.h>
 #include <bpf/bpf.h>
+#include "util/bpf_map.h"
 #include "builtin.h"
 #include "util/cgroup.h"
 #include "util/color.h"
@@ -87,6 +88,9 @@ struct trace {
 					  *augmented;
 		}		events;
 	} syscalls;
+	struct {
+		struct bpf_map *map;
+	} dump;
 	struct record_opts	opts;
 	struct perf_evlist	*evlist;
 	struct machine		*host;
@@ -2997,6 +3001,9 @@ static int trace__run(struct trace *trace, int argc, const char **argv)
 	if (err < 0)
 		goto out_error_apply_filters;
 
+	if (trace->dump.map)
+		bpf_map__fprintf(trace->dump.map, trace->output);
+
 	err = perf_evlist__mmap(evlist, trace->opts.mmap_pages);
 	if (err < 0)
 		goto out_error_mmap;
@@ -3686,6 +3693,7 @@ int cmd_trace(int argc, const char **argv)
 		.max_stack = UINT_MAX,
 		.max_events = ULONG_MAX,
 	};
+	const char *map_dump_str = NULL;
 	const char *output_name = NULL;
 	const struct option trace_options[] = {
 	OPT_CALLBACK('e', "event", &trace, "event",
@@ -3718,6 +3726,9 @@ int cmd_trace(int argc, const char **argv)
 	OPT_CALLBACK(0, "duration", &trace, "float",
 		     "show only events with duration > N.M ms",
 		     trace__set_duration),
+#ifdef HAVE_LIBBPF_SUPPORT
+	OPT_STRING(0, "map-dump", &map_dump_str, "BPF map", "BPF map to periodically dump"),
+#endif
 	OPT_BOOLEAN(0, "sched", &trace.sched, "show blocking scheduler events"),
 	OPT_INCR('v', "verbose", &verbose, "be more verbose"),
 	OPT_BOOLEAN('T', "time", &trace.full_time,
@@ -3812,6 +3823,14 @@ int cmd_trace(int argc, const char **argv)
 
 	err = -1;
 
+	if (map_dump_str) {
+		trace.dump.map = bpf__find_map_by_name(map_dump_str);
+		if (trace.dump.map == NULL) {
+			pr_err("ERROR: BPF map \"%s\" not found\n", map_dump_str);
+			goto out;
+		}
+	}
+
 	if (trace.trace_pgfaults) {
 		trace.opts.sample_address = true;
 		trace.opts.sample_time = true;
Signature · 1 line
2.19.1
author maintainer
#12 lore raw

PATCH 12/17 perf session: Don't report zero period samples for slave events

In reply to Arnaldo Carvalho de Melo

From: Jiri Olsa <[email protected]>

There's no reason to deliver a sample with zero period. It means there was no value for slave event since its last group leader sample.

---
 tools/perf/util/session.c | 7 +++++++
 1 file changed, 7 insertions(+)
tools/perf/util/session.c +7 -0
index 18fb9c8cbf9c..c764bbc91009 100644
@@ -1202,6 +1202,13 @@ static int deliver_sample_value(struct perf_evlist *evlist,
 		return 0;
 	}
 
+	/*
+	 * There's no reason to deliver sample
+	 * for zero period, bail out.
+	 */
+	if (!sample->period)
+		return 0;
+
 	return tool->sample(tool, event, sample, sid->evsel, machine);
 }
 
Signature · 1 line
2.19.1
author maintainer
#13 lore raw

PATCH 13/17 perf evsel: Force sample_type for slave events

In reply to Arnaldo Carvalho de Melo

From: Jiri Olsa <[email protected]>

Force sample_type setup for slave events in group leader sessions.

We don't get sample for slave events, we make them when delivering group leader sample. Set the slave event to follow the master sample_type to ease up report.

---
 tools/perf/util/evsel.c | 8 ++++++++
 1 file changed, 8 insertions(+)
tools/perf/util/evsel.c +8 -0
index 684c893ca6bc..dfe2958e6287 100644
@@ -956,6 +956,14 @@ void perf_evsel__config(struct perf_evsel *evsel, struct record_opts *opts,
 		attr->sample_freq    = 0;
 		attr->sample_period  = 0;
 		attr->write_backward = 0;
+
+		/*
+		 * We don't get sample for slave events, we make them
+		 * when delivering group leader sample. Set the slave
+		 * event to follow the master sample_type to ease up
+		 * report.
+		 */
+		attr->sample_type = leader->attr.sample_type;
 	}
 
 	if (opts->no_samples)
Signature · 1 line
2.19.1
author maintainer
#14 lore raw

PATCH 14/17 perf script: Allow +- operator for type specific fields option

In reply to Arnaldo Carvalho de Melo

From: Jiri Olsa <[email protected]>

Add support to add/remove fields for specific event types in -F option. It's now possible to use '+-' after event type, like:

  # cat > test.c
  #include <stdio.h>
  int main(void)
  {
     printf("Hello world\n");
     while(1) {}
  }
  ^D
  # gcc -g -o test test.c
  # perf probe -x test 'test.c:5'
  # perf record -e '{cpu/cpu-cycles,period=10000/,probe_test:main}:S' ./test
  ...
  # perf script -Ftrace:+period,-cpu
            test  3859 396291.117343:      10275 cpu/cpu-cycles,period=10000/:      7f..
            test  3859 396291.118234:      11041 cpu/cpu-cycles,period=10000/:  ffffff..
            test  3859 396291.118234:          1              probe_test:main:
            test  3859 396291.118248:       8668 cpu/cpu-cycles,period=10000/:  ffffff..
            test  3859 396291.118263:      10139 cpu/cpu-cycles,period=10000/:  ffffff..

Committer testing:

Couldn't make the test above work, but tested it with:

  # perf probe -x hello main
  Added new event:
    probe_hello:main     (on main in /home/acme/c/hello)

You can now use it in all perf tools, such as:

perf record -e probe_hello:main -aR sleep 1

  # perf record -e probe_hello:main ./hello
  hello, world
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 0.025 MB perf.data (1 samples) ]
  # perf script
           hello 21454 [002] 254116.874005: probe_hello:main: (401126)
  #
  # perf script -Ftrace:+period,-cpu
           hello 21454 254116.874005:          1 probe_hello:main: (401126)
---
 tools/perf/Documentation/perf-script.txt | 6 ++++++
 tools/perf/builtin-script.c              | 8 ++++++++
 2 files changed, 14 insertions(+)
tools/perf/Documentation/perf-script.txt +6 -0
index 9e4def08d569..2e19fd7ffe35 100644
@@ -159,6 +159,12 @@ OPTIONS
 	the override, and the result of the above is that only S/W and H/W
 	events are displayed with the given fields.
 
+	It's possible tp add/remove fields only for specific event type:
+
+		-Fsw:-cpu,-period
+
+	removes cpu and period from software events.
+
 	For the 'wildcard' option if a user selected field is invalid for an
 	event type, a message is displayed to the user that the option is
 	ignored for that type. For example:
tools/perf/builtin-script.c +8 -0
index 8d5fe092525c..373ea151dc60 100644
@@ -2560,6 +2560,10 @@ static int parse_output_fields(const struct option *opt __maybe_unused,
 			pr_warning("Overriding previous field request for %s events.\n",
 				   event_type(type));
 
+		/* Don't override defaults for +- */
+		if (strchr(tok, '+') || strchr(tok, '-'))
+			goto parse;
+
 		output[type].fields = 0;
 		output[type].user_set = true;
 		output[type].wildcard_set = false;
@@ -2644,6 +2648,10 @@ static int parse_output_fields(const struct option *opt __maybe_unused,
 				rc = -EINVAL;
 				goto out;
 			}
+			if (change == REMOVE)
+				output[type].fields &= ~all_output_options[i].field;
+			else
+				output[type].fields |= all_output_options[i].field;
 			output[type].user_set = true;
 			output[type].wildcard_set = true;
 		}
Signature · 1 line
2.19.1
author maintainer
#15 lore raw

PATCH 15/17 perf bpf-event: Add missing new line into pr_debug call

In reply to Arnaldo Carvalho de Melo

From: Jiri Olsa <[email protected]>

Add a missing new line into pr_debug call in perf_event__synthesize_bpf_events(), so that the error message does not screw the verbose output.

---
 tools/perf/util/bpf-event.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/util/bpf-event.c +1 -1
index 62dda96b0096..028c8ec1f62a 100644
@@ -233,7 +233,7 @@ int perf_event__synthesize_bpf_events(struct perf_tool *tool,
 				err = 0;
 				break;
 			}
-			pr_debug("%s: can't get next program: %s%s",
+			pr_debug("%s: can't get next program: %s%s\n",
 				 __func__, strerror(errno),
 				 errno == EINVAL ? " -- kernel too old?" : "");
 			/* don't report error on old kernel or EPERM  */
Signature · 1 line
2.19.1
author maintainer
#17 lore raw

PATCH 16/17 perf cpumap: Increase debug level for cpu_map__snprint verbose output

In reply to Arnaldo Carvalho de Melo

From: Jiri Olsa <[email protected]>

So it does not screw up single -v verbose output.

---
 tools/perf/util/cpumap.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/util/cpumap.c +1 -1
index 0bbc3feb0894..0b599229bc7e 100644
@@ -681,7 +681,7 @@ size_t cpu_map__snprint(struct cpu_map *map, char *buf, size_t size)
 
 #undef COMMA
 
-	pr_debug("cpumask list: %s\n", buf);
+	pr_debug2("cpumask list: %s\n", buf);
 	return ret;
 }
 
Signature · 1 line
2.19.1
author maintainer
#18 lore raw

PATCH 17/17 perf tools: Make rm_rf() remove single file

In reply to Arnaldo Carvalho de Melo

From: Jiri Olsa <[email protected]>

Let rm_rf() remove a file if it's provided by path, not just directories.

---
 tools/perf/util/util.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)
tools/perf/util/util.c +13 -3
index 320b0fef249a..3ee410fc047a 100644
@@ -120,16 +120,26 @@ int mkdir_p(char *path, mode_t mode)
 int rm_rf(const char *path)
 {
 	DIR *dir;
-	int ret = 0;
+	int ret;
 	struct dirent *d;
 	char namebuf[PATH_MAX];
+	struct stat statbuf;
 
+	/* Do not fail if there's no file. */
+	ret = lstat(path, &statbuf);
+	if (ret)
+		return 0;
+
+	/* Try to remove any file we get. */
+	if (!(statbuf.st_mode & S_IFDIR))
+		return unlink(path);
+
+	/* We have directory in path. */
 	dir = opendir(path);
 	if (dir == NULL)
-		return 0;
+		return -1;
 
 	while ((d = readdir(dir)) != NULL && !ret) {
-		struct stat statbuf;
 
 		if (!strcmp(d->d_name, ".") || !strcmp(d->d_name, ".."))
 			continue;
Signature · 1 line
2.19.1
Acked-by on v1 15/17
#16 lore raw

Re: [PATCH 15/17] perf bpf-event: Add missing new line into pr_debug call

In reply to Arnaldo Carvalho de Melo
Quoted 9 lines · Arnaldo Carvalho de Melo

From: Jiri Olsa <[email protected]>

Add a missing new line into pr_debug call in perf_event__synthesize_bpf_events(), so that the error message does not screw the verbose output.

Quoted 8 lines

Thanks for fixing this.

Song

Quoted 20 lines

---
tools/perf/util/bpf-event.c | 2 +-

1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/util/bpf-event.c +1 -1
index 62dda96b0096..028c8ec1f62a 100644
@@ -233,7 +233,7 @@ int perf_event__synthesize_bpf_events(struct perf_tool *tool,
 				err = 0;
 				break;
 			}
-			pr_debug("%s: can't get next program: %s%s",
+			pr_debug("%s: can't get next program: %s%s\n",
 				 __func__, strerror(errno),
 				 errno == EINVAL ? " -- kernel too old?" : "");
 			/* don't report error on old kernel or EPERM  */
Signature · 1 line
2.19.1
#19 lore raw
In reply to Arnaldo Carvalho de Melo
Quoted 130 lines · Arnaldo Carvalho de Melo

Hi Ingo,

Please consider pulling,

- Arnaldo

Test results at the end of this message, as usual.

The following changes since commit 43f4e6279f05eefac058a3524e184cecae463bfe:

Merge tag 'perf-core-for-mingo-5.1-20190214' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core (2019-02-15 10:19:11 +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-5.1-20190220

for you to fetch changes up to b4409ae112caa6315f6ee678e953b9fc93e6919c:

perf tools: Make rm_rf() remove single file (2019-02-20 17:09:28 -0300)

----------------------------------------------------------------
perf/core improvements and fixes:

perf report:

He Kuang:

- Don't shadow inlined symbol with different addr range.

perf script:

Jiri Olsa:

- Allow +- operator to ask for -F to add/remove fields to the default set, for instance to ask for the removal of the 'cpu' field in tracepoint events, adding 'period' to that kind of events, etc.

perf test:

Thomas Richter:

- Fix scheduler tracepoint signedness of COMM fields failure of 'evsel-tp-sched' test on s390 and other arches.

Tommi Rantala:

- Skip trace+probe_vfs_getname.sh when 'perf trace' is not built.

perf trace:

Arnaldo Carvalho de Melo:

- Add initial BPF map dumper, initially just for the current, minimal needs of the augmented_raw_syscalls BPF example used to collect pointer args payloads that uses BPF maps for pid and syscall filtering, but will in time have features similar to 'perf stat' --interval-print, --interval-clear, ways to signal from a BPF event that a specific map (or range of that map) should be printed, optionally as a histogram, etc.

General:

Jiri Olsa:

- Add cpu and numa topologies classes for further reuse, fixing some issues in the process.

- Fixup some warnings and debug levels.

- Make rm_rf() remove single file, not just directories.

Documentation:

Jonas Rabenstein:

- Fix HEADER_CMDLINE description in perf.data documentation.

- Fix documentation of the Flags section in perf.data.

----------------------------------------------------------------
Arnaldo Carvalho de Melo (2):
perf bpf: Add bpf_map dumper
perf trace: Allow dumping a BPF map after setting up BPF events

He Kuang (1):
perf report: Don't shadow inlined symbol with different addr range

Jiri Olsa (10):
perf header: Fix wrong node write in NUMA_TOPOLOGY feature perf tools: Add cpu_topology object
perf tools: Add numa_topology object
perf tools: Use sysfs__mountpoint() when reading cpu topology perf session: Don't report zero period samples for slave events perf evsel: Force sample_type for slave events
perf script: Allow +- operator for type specific fields option perf bpf-event: Add missing new line into pr_debug call perf cpumap: Increase debug level for cpu_map__snprint verbose output perf tools: Make rm_rf() remove single file

Jonas Rabenstein (2):
perf doc: Fix HEADER_CMDLINE description in perf.data documentation perf doc: Fix documentation of the Flags section in perf.data

Thomas Richter (1):
perf test: Fix failure of 'evsel-tp-sched' test on s390

Tommi Rantala (1):
perf tests shell: Skip trace+probe_vfs_getname.sh if built without trace support

 tools/perf/Documentation/perf-script.txt           |   6 +
 tools/perf/Documentation/perf-trace.txt            |   8 +
 tools/perf/Documentation/perf.data-file-format.txt |  11 +-
 tools/perf/builtin-script.c                        |   8 +
 tools/perf/builtin-trace.c                         |  19 ++
 tools/perf/tests/evsel-tp-sched.c                  |   6 +-
 tools/perf/tests/shell/lib/probe.sh                |   5 +
 tools/perf/tests/shell/trace+probe_vfs_getname.sh  |   1 +
 tools/perf/util/Build                              |   2 +
 tools/perf/util/bpf-event.c                        |   2 +-
 tools/perf/util/bpf_map.c                          |  72 ++++++
 tools/perf/util/bpf_map.h                          |  22 ++
 tools/perf/util/cpumap.c                           |   2 +-
 tools/perf/util/cputopo.c                          | 277 +++++++++++++++++++++
 tools/perf/util/cputopo.h                          |  33 +++
 tools/perf/util/evsel.c                            |   8 +
 tools/perf/util/header.c                           | 269 +++-----------------
 tools/perf/util/session.c                          |   7 +
Quoted 8 lines
 tools/perf/util/sort.c                             |  10 +-
 tools/perf/util/srcline.c                          |   2 +-
 tools/perf/util/util.c                             |  16 +-
 21 files changed, 530 insertions(+), 256 deletions(-)
 create mode 100644 tools/perf/util/bpf_map.c
 create mode 100644 tools/perf/util/bpf_map.h
 create mode 100644 tools/perf/util/cputopo.c
 create mode 100644 tools/perf/util/cputopo.h

Pulled, thanks a lot Arnaldo!

Ingo