perf/core improvements and fixes
Subsystems (3)
Arnaldo Carvalho de Melo
maintainerHi Ingo/Thomas,
Please consider pulling,
Best regards,
- Arnaldo
Test results at the end of this message, as usual.
The following changes since commit 39b656ee9f2ce41eb969c86525f9a2a63fefac5b:
Merge tag 'perf-core-for-mingo-5.5-20191011' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core (2019-10-15 07:19:55 +0200)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git tags/perf-core-for-mingo-5.5-20191021
for you to fetch changes up to 27198a893ba074407e7a87e346252b3e6fab454f:
perf trace: Use STUL_STRARRAY_FLAGS with mmap (2019-10-19 15:35:02 -0300)
----------------------------------------------------------------
perf/core improvements and fixes:
perf trace:
- Add syscall failure stats to -s/--summary and -S/--with-summary, also works in combination with specifying just a set of syscalls, see below first with -s/--summary, then with -S/--with-summary just for the syscalls we saw failing with -s:
# perf trace -s sleep 1
Summary of events:
sleep (16218), 80 events, 93.0%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
----------- ----- ------ -------- -------- -------- -------- ------
nanosleep 1 0 1000.091 1000.091 1000.091 1000.091 0.00%
mmap 8 0 0.045 0.005 0.006 0.008 7.09%
mprotect 4 0 0.028 0.005 0.007 0.009 11.38%
openat 3 0 0.021 0.005 0.007 0.009 14.07%
munmap 1 0 0.017 0.017 0.017 0.017 0.00%
brk 4 0 0.010 0.001 0.002 0.004 23.15%
read 4 0 0.009 0.002 0.002 0.003 8.13%
close 5 0 0.008 0.001 0.002 0.002 10.83%
fstat 3 0 0.006 0.002 0.002 0.002 6.97%
access 1 1 0.006 0.006 0.006 0.006 0.00%
lseek 3 0 0.005 0.001 0.002 0.002 7.37%
arch_prctl 2 1 0.004 0.001 0.002 0.002 17.64%
execve 1 0 0.000 0.000 0.000 0.000 0.00%
# perf trace -e access,arch_prctl -S sleep 1
0.000 ( 0.006 ms): sleep/19503 arch_prctl(option: 0x3001, arg2: 0x7fff165996b0) = -1 EINVAL (Invalid argument)
0.024 ( 0.006 ms): sleep/19503 access(filename: 0x2177e510, mode: R) = -1 ENOENT (No such file or directory)
0.136 ( 0.002 ms): sleep/19503 arch_prctl(option: SET_FS, arg2: 0x7f9421737580) = 0
Summary of events:
sleep (19503), 6 events, 50.0%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
---------- ----- ------ ------ ------ ------ ------ ------
arch_prctl 2 1 0.008 0.002 0.004 0.006 57.22%
access 1 1 0.006 0.006 0.006 0.006 0.00%
#- Introduce --errno-summary, to drill down a bit more in the errno stats:
# perf trace --errno-summary -e access,arch_prctl -S sleep 1
0.000 ( 0.006 ms): sleep/5587 arch_prctl(option: 0x3001, arg2: 0x7ffd6ba6aa00) = -1 EINVAL (Invalid argument)
0.028 ( 0.007 ms): sleep/5587 access(filename: 0xb83d9510, mode: R) = -1 ENOENT (No such file or directory)
0.172 ( 0.003 ms): sleep/5587 arch_prctl(option: SET_FS, arg2: 0x7f45b8392580) = 0 Summary of events:
sleep (5587), 6 events, 50.0%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
---------- ----- ------ ------ ------ ------ ------ ------
arch_prctl 2 1 0.009 0.003 0.005 0.006 38.90%
EINVAL: 1
access 1 1 0.007 0.007 0.007 0.007 0.00%
ENOENT: 1
#- Filter own pid to avoid a feedback look in 'perf trace record -a'
- Add the glue for the auto generated x86 IRQ vector array.
- Show error message when not finding a field used in a filter expression
# perf trace --max-events=4 -e syscalls:sys_enter_write --filter="cnt>32767"
Failed to set filter "(cnt>32767) && (common_pid != 19938 && common_pid != 8922)" on event syscalls:sys_enter_write with 22 (Invalid argument)
#
# perf trace --max-events=4 -e syscalls:sys_enter_write --filter="count>32767"
0.000 python3.5/17535 syscalls:sys_enter_write(fd: 3, buf: 0x564b0dc53600, count: 172086)
12.641 python3.5.post/17535 syscalls:sys_enter_write(fd: 3, buf: 0x564b0db63660, count: 75994)
27.738 python3.5.post/17535 syscalls:sys_enter_write(fd: 3, buf: 0x564b0db4b1e0, count: 41635)
136.070 python3.5.post/17535 syscalls:sys_enter_write(fd: 3, buf: 0x564b0dbab510, count: 62232)
#- Add a generator for x86's IRQ vectors -> strings
- Introduce stroul() (string -> number) methods for the strarray and
strarrays classes, also strtoul_flags, allowing to go from both strings
and or-ed strings to numbers, allowing things like:
# perf trace -e syscalls:sys_enter_mmap --filter="flags==DENYWRITE|PRIVATE|FIXED" sleep 1
0.000 sleep/22588 syscalls:sys_enter_mmap(addr: 0x7f42d2aa5000, len: 1363968, prot: READ|EXEC, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x22000)
0.011 sleep/22588 syscalls:sys_enter_mmap(addr: 0x7f42d2bf2000, len: 311296, prot: READ, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x16f000)
0.015 sleep/22588 syscalls:sys_enter_mmap(addr: 0x7f42d2c3f000, len: 24576, prot: READ|WRITE, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x1bb000)
#Allowing to narrow down from the complete set of mmap calls for that workload:
# perf trace -e syscalls:sys_enter_mmap sleep 1
0.000 sleep/22695 syscalls:sys_enter_mmap(len: 134773, prot: READ, flags: PRIVATE, fd: 3)
0.041 sleep/22695 syscalls:sys_enter_mmap(len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS)
0.053 sleep/22695 syscalls:sys_enter_mmap(len: 1857472, prot: READ, flags: PRIVATE|DENYWRITE, fd: 3)
0.069 sleep/22695 syscalls:sys_enter_mmap(addr: 0x7fd23ffb6000, len: 1363968, prot: READ|EXEC, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x22000)
0.077 sleep/22695 syscalls:sys_enter_mmap(addr: 0x7fd240103000, len: 311296, prot: READ, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x16f000)
0.083 sleep/22695 syscalls:sys_enter_mmap(addr: 0x7fd240150000, len: 24576, prot: READ|WRITE, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x1bb000)
0.095 sleep/22695 syscalls:sys_enter_mmap(addr: 0x7fd240156000, len: 14272, prot: READ|WRITE, flags: PRIVATE|FIXED|ANONYMOUS)
0.339 sleep/22695 syscalls:sys_enter_mmap(len: 217750512, prot: READ, flags: PRIVATE, fd: 3)
#Works with all targets, so, for system wide, looking at who calls mmap with flags set to just "PRIVATE":
# perf trace --max-events=5 -e syscalls:sys_enter_mmap --filter="flags==PRIVATE"
0.000 pool/2242 syscalls:sys_enter_mmap(len: 756, prot: READ, flags: PRIVATE, fd: 14)
0.050 pool/2242 syscalls:sys_enter_mmap(len: 756, prot: READ, flags: PRIVATE, fd: 14)
0.062 pool/2242 syscalls:sys_enter_mmap(len: 756, prot: READ, flags: PRIVATE, fd: 14)
0.145 goa-identity-s/2240 syscalls:sys_enter_mmap(len: 756, prot: READ, flags: PRIVATE, fd: 18)
0.183 goa-identity-s/2240 syscalls:sys_enter_mmap(len: 756, prot: READ, flags: PRIVATE, fd: 18)
# # perf trace --max-events=2 -e syscalls:sys_enter_lseek --filter="whence==SET && offset != 0"
0.000 Cache2 I/O/12047 syscalls:sys_enter_lseek(fd: 277, offset: 43, whence: SET)
1142.070 mozStorage #5/12302 syscalls:sys_enter_lseek(fd: 44</home/acme/.mozilla/firefox/ina67tev.default/cookies.sqlite-wal>, offset: 393536, whence: SET)
#perf annotate:
- Fix objdump --no-show-raw-insn flag to work with goth gcc and clang.
- Streamline objdump execution, preserving the right error codes for better reporting to user.
perf report:
- Add warning when libunwind not compiled in.
perf stat:
Jin Yao:
- Support --all-kernel/--all-user, to match options available in 'perf record', asking that all the events specified work just with kernel or user events.
perf list:
Jin Yao:
- Hide deprecated events by default, allow showing them with --deprecated.
libbperf:
Jiri Olsa:
- Allow to build with -ltcmalloc.
- Finish mmap interface, getting more stuff from tools/perf while adding abstractions to avoid pulling too much stuff, to get libperf to grow as tools needs things like auxtrace, etc.
perf scripting engines:
Steven Rostedt (VMware):
- Iterate on tep event arrays directly, fixing script generation with '-g python' when having multiple tracepoints in a perf.data file.
core:
- Allow to build with -ltcmalloc.
perf test:
Leo Yan:
- Report failure for mmap events.
- Avoid infinite loop for task exit case.
- Remove needless headers for bp_account test.
- Add dedicated checking helper is_supported().
- Disable bp_signal testing for arm64.
Vendor events:
arm64:
John Garry:
- Fix Hisi hip08 DDRC PMU eventname.
- Add some missing events for Hisi hip08 DDRC, L3C and HHA PMUs.
----------------------------------------------------------------
Andi Kleen (2):
perf script: Fix --reltime with --time
perf evlist: Fix fix for freed id arrays
Arnaldo Carvalho de Melo (25):
perf trace: Add syscall failure stats to -s/--summary and -S/--with-summary perf trace: Introduce --errno-summary
perf string: Export asprintf__tp_filter_pids()
perf trace: Filter own pid to avoid a feedback look in 'perf trace record -a' perf trace: Support tracepoint dynamic char arrays
tools arch x86: Grab a copy of the file containing the IRQ vector defines
libbeauty: Add a generator for x86's IRQ vectors -> strings
libbeauty: Hook up the x86 irq_vectors table generator
libbeauty: Add a strarray__scnprintf_suffix() method
perf trace beauty: Add the glue for the autogenerated x86 IRQ vector array perf trace: Hook the 'vec' tracepoint argument with the x86 IRQ vectors scnprintf/strtoul perf trace: Show error message when not finding a field used in a filter expression perf trace: Introduce accessors to trace specific evsel->priv
perf trace: Hide evsel->access further, simplify code
perf trace: Introduce 'struct evsel__trace' for evsel->priv needs
perf trace: Initialize evsel_trace->fmt for syscalls:sys_enter_* tracepoints
libbeauty: Introduce syscall_arg__strtoul_strarray()
perf trace: Honour --max-events in processing syscalls:sys_enter_*
perf trace: Pass a syscall_arg to syscall_arg_fmt->strtoul()
libbeauty: Introduce syscall_arg__strtoul_strarrays()
perf trace: Use strtoul for the fcntl 'cmd' argument
libbeauty: Make the mmap_flags strarray visible outside of its beautifier
libbeauty: Introduce strarray__strtoul_flags()
perf trace: Wire up strarray__strtoul_flags()
perf trace: Use STUL_STRARRAY_FLAGS with mmap
Ian Rogers (5):
perf annotate: Avoid reallocation in objdump parsing perf annotate: Use libsubcmd's run-command.h to fork objdump perf annotate: Don't pipe objdump output through 'grep' command perf annotate: Don't pipe objdump output through 'expand' command perf annotate: Fix objdump --no-show-raw-insn flag
Jin Yao (3):
perf report: Add warning when libunwind not compiled in perf stat: Support --all-kernel/--all-user perf list: Hide deprecated events by default
Jiri Olsa (10):
perf tools: Allow to build with -ltcmalloc
libperf: Introduce perf_evlist__for_each_mmap()
libperf: Move mmap allocation to perf_evlist__mmap_ops::get
libperf: Move mask setup to perf_evlist__mmap_ops()
libperf: Link static tests with libapi.a
libperf: Add tests_mmap_thread test
libperf: Add tests_mmap_cpus test
libperf: Keep count of failed tests
libperf: Do not export perf_evsel__init()/perf_evlist__init()
libperf: Add pr_err() macro
John Garry (4):
perf vendor events arm64: Fix Hisi hip08 DDRC PMU eventname perf vendor events arm64: Add some missing events for Hisi hip08 DDRC PMU perf vendor events arm64: Add some missing events for Hisi hip08 L3C PMU perf vendor events arm64: Add some missing events for Hisi hip08 HHA PMU
Leo Yan (5):
perf test: Report failure for mmap events
perf test: Avoid infinite loop for task exit case
perf tests: Remove needless headers for bp_account
perf tests bp_account: Add dedicated checking helper is_supported() perf tests: Disable bp_signal testing for arm64
Steven Rostedt (VMware) (2):
perf scripting engines: Iterate on tep event arrays directly perf tools: Remove unused trace_find_next_event()
Thomas Richter (1):
perf jvmti: Link against tools/lib/ctype.h to have weak strlcpy()
tools/arch/x86/include/asm/irq_vectors.h | 146 +++++++ tools/perf/Documentation/perf-list.txt | 3 + tools/perf/Documentation/perf-stat.txt | 6 + tools/perf/Documentation/perf-trace.txt | 4 + tools/perf/Makefile.config | 5 + tools/perf/Makefile.perf | 10 + tools/perf/builtin-list.c | 14 +- tools/perf/builtin-report.c | 7 + tools/perf/builtin-script.c | 5 +- tools/perf/builtin-stat.c | 6 + tools/perf/builtin-trace.c | 420 ++++++++++++++++----- tools/perf/check-headers.sh | 1 + tools/perf/jvmti/Build | 6 +- tools/perf/lib/Makefile | 1 + tools/perf/lib/evlist.c | 71 +++- tools/perf/lib/include/internal/evlist.h | 3 + tools/perf/lib/include/internal/evsel.h | 1 + tools/perf/lib/include/internal/mmap.h | 5 +- tools/perf/lib/include/internal/tests.h | 20 +- tools/perf/lib/include/perf/core.h | 1 + tools/perf/lib/include/perf/evlist.h | 10 +- tools/perf/lib/include/perf/evsel.h | 2 - tools/perf/lib/internal.h | 3 + tools/perf/lib/libperf.map | 3 +- tools/perf/lib/mmap.c | 6 +- tools/perf/lib/tests/Makefile | 6 +- tools/perf/lib/tests/test-cpumap.c | 2 +- tools/perf/lib/tests/test-evlist.c | 219 ++++++++++- tools/perf/lib/tests/test-evsel.c | 2 +- tools/perf/lib/tests/test-threadmap.c | 2 +- .../arch/arm64/hisilicon/hip08/uncore-ddrc.json | 16 +- .../arch/arm64/hisilicon/hip08/uncore-hha.json | 23 +- .../arch/arm64/hisilicon/hip08/uncore-l3c.json | 56 +++ tools/perf/pmu-events/jevents.c | 26 +- tools/perf/pmu-events/jevents.h | 3 +- tools/perf/pmu-events/pmu-events.h | 1 + tools/perf/tests/bp_account.c | 20 +- tools/perf/tests/bp_signal.c | 15 +- tools/perf/tests/builtin-test.c | 2 +- tools/perf/tests/task-exit.c | 9 + tools/perf/tests/tests.h | 1 + tools/perf/trace/beauty/beauty.h | 19 + tools/perf/trace/beauty/mmap.c | 4 +- tools/perf/trace/beauty/tracepoints/Build | 1 + .../trace/beauty/tracepoints/x86_irq_vectors.c | 29 ++ .../trace/beauty/tracepoints/x86_irq_vectors.sh | 27 ++ tools/perf/util/annotate.c | 196 ++++++---- tools/perf/util/evlist.c | 34 +- tools/perf/util/parse-events.c | 4 +- tools/perf/util/parse-events.h | 2 +- tools/perf/util/pmu.c | 17 +- tools/perf/util/pmu.h | 4 +- .../perf/util/scripting-engines/trace-event-perl.c | 8 +- .../util/scripting-engines/trace-event-python.c | 9 +- tools/perf/util/stat.c | 10 + tools/perf/util/stat.h | 2 + tools/perf/util/string2.h | 3 + tools/perf/util/time-utils.c | 27 +- tools/perf/util/time-utils.h | 5 + tools/perf/util/trace-event-parse.c | 31 -- tools/perf/util/trace-event.h | 2 - 61 files changed, 1307 insertions(+), 289 deletions(-) create mode 100644 tools/arch/x86/include/asm/irq_vectors.h create mode 100644 tools/perf/trace/beauty/tracepoints/x86_irq_vectors.c create mode 100755 tools/perf/trace/beauty/tracepoints/x86_irq_vectors.sh
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.
Clearlinux is failing when building with libpython, but that is not a perf regression, will try to remove one compiler warning that is causing the problem when building some of the glue code files in the python files, outside perf.
# export PERF_TARBALL=http://192.168.124.1/perf/perf-5.4.0-rc3.tar.xz # dm 1 alpine:3.4 : Ok gcc (Alpine 5.3.0) 5.3.0, clang version 3.8.0 (tags/RELEASE_380/final) 2 alpine:3.5 : Ok gcc (Alpine 6.2.1) 6.2.1 20160822, clang version 3.8.1 (tags/RELEASE_381/final) 3 alpine:3.6 : Ok gcc (Alpine 6.3.0) 6.3.0, clang version 4.0.0 (tags/RELEASE_400/final) 4 alpine:3.7 : Ok gcc (Alpine 6.4.0) 6.4.0, Alpine clang version 5.0.0 (tags/RELEASE_500/final) (based on LLVM 5.0.0) 5 alpine:3.8 : Ok gcc (Alpine 6.4.0) 6.4.0, Alpine clang version 5.0.1 (tags/RELEASE_501/final) (based on LLVM 5.0.1) 6 alpine:3.9 : Ok gcc (Alpine 8.3.0) 8.3.0, Alpine clang version 5.0.1 (tags/RELEASE_502/final) (based on LLVM 5.0.1) 7 alpine:3.10 : Ok gcc (Alpine 8.3.0) 8.3.0, Alpine clang version 8.0.0 (tags/RELEASE_800/final) (based on LLVM 8.0.0) 8 alpine:edge : Ok gcc (Alpine 9.2.0) 9.2.0, Alpine clang version 8.0.1 (tags/RELEASE_801/final) (based on LLVM 8.0.1) 9 amazonlinux:1 : Ok gcc (GCC) 7.2.1 20170915 (Red Hat 7.2.1-2), clang version 3.6.2 (tags/RELEASE_362/final) 10 amazonlinux:2 : Ok gcc (GCC) 7.3.1 20180303 (Red Hat 7.3.1-5), clang version 7.0.1 (Amazon Linux 2 7.0.1-1.amzn2.0.2) 11 android-ndk:r12b-arm : Ok arm-linux-androideabi-gcc (GCC) 4.9.x 20150123 (prerelease) 12 android-ndk:r15c-arm : Ok arm-linux-androideabi-gcc (GCC) 4.9.x 20150123 (prerelease) 13 centos:5 : Ok gcc (GCC) 4.1.2 20080704 (Red Hat 4.1.2-55) 14 centos:6 : Ok gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-23) 15 centos:7 : Ok gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-39), clang version 3.4.2 (tags/RELEASE_34/dot2-final) 16 centos:8 : Ok gcc (GCC) 8.2.1 20180905 (Red Hat 8.2.1-3), clang version 7.0.1 (tags/RELEASE_701/final) 17 clearlinux:latest : Ok gcc (Clear Linux OS for Intel Architecture) 9.2.1 20190930 gcc-9-branch@276275, clang version 8.0.0 (tags/RELEASE_800/final) 18 debian:8 : Ok gcc (Debian 4.9.2-10+deb8u2) 4.9.2, Debian clang version 3.5.0-10 (tags/RELEASE_350/final) (based on LLVM 3.5.0) 19 debian:9 : Ok gcc (Debian 6.3.0-18+deb9u1) 6.3.0 20170516, clang version 3.8.1-24 (tags/RELEASE_381/final) 20 debian:10 : Ok gcc (Debian 8.3.0-6) 8.3.0, clang version 7.0.1-8 (tags/RELEASE_701/final) 21 debian:experimental : Ok gcc (Debian 9.2.1-8) 9.2.1 20190909, clang version 8.0.1-3+b1 (tags/RELEASE_801/final) 22 debian:experimental-x-arm64 : Ok aarch64-linux-gnu-gcc (Debian 8.3.0-19) 8.3.0 23 debian:experimental-x-mips : Ok mips-linux-gnu-gcc (Debian 8.3.0-19) 8.3.0 24 debian:experimental-x-mips64 : Ok mips64-linux-gnuabi64-gcc (Debian 8.3.0-19) 8.3.0 25 debian:experimental-x-mipsel : Ok mipsel-linux-gnu-gcc (Debian 8.3.0-19) 8.3.0 26 fedora:20 : Ok gcc (GCC) 4.8.3 20140911 (Red Hat 4.8.3-7), clang version 3.4.2 (tags/RELEASE_34/dot2-final) 27 fedora:22 : Ok gcc (GCC) 5.3.1 20160406 (Red Hat 5.3.1-6), clang version 3.5.0 (tags/RELEASE_350/final) 28 fedora:23 : Ok gcc (GCC) 5.3.1 20160406 (Red Hat 5.3.1-6), clang version 3.7.0 (tags/RELEASE_370/final) 29 fedora:24 : Ok gcc (GCC) 6.3.1 20161221 (Red Hat 6.3.1-1), clang version 3.8.1 (tags/RELEASE_381/final) 30 fedora:24-x-ARC-uClibc : Ok arc-linux-gcc (ARCompact ISA Linux uClibc toolchain 2017.09-rc2) 7.1.1 20170710 31 fedora:25 : Ok gcc (GCC) 6.4.1 20170727 (Red Hat 6.4.1-1), clang version 3.9.1 (tags/RELEASE_391/final) 32 fedora:26 : Ok gcc (GCC) 7.3.1 20180130 (Red Hat 7.3.1-2), clang version 4.0.1 (tags/RELEASE_401/final) 33 fedora:27 : Ok gcc (GCC) 7.3.1 20180712 (Red Hat 7.3.1-6), clang version 5.0.2 (tags/RELEASE_502/final) 34 fedora:28 : Ok gcc (GCC) 8.3.1 20190223 (Red Hat 8.3.1-2), clang version 6.0.1 (tags/RELEASE_601/final) 35 fedora:29 : Ok gcc (GCC) 8.3.1 20190223 (Red Hat 8.3.1-2), clang version 7.0.1 (Fedora 7.0.1-6.fc29) 36 fedora:30 : Ok gcc (GCC) 9.2.1 20190827 (Red Hat 9.2.1-1), clang version 8.0.0 (Fedora 8.0.0-3.fc30) 37 fedora:30-x-ARC-glibc : Ok arc-linux-gcc (ARC HS GNU/Linux glibc toolchain 2019.03-rc1) 8.3.1 20190225 38 fedora:30-x-ARC-uClibc : Ok arc-linux-gcc (ARCv2 ISA Linux uClibc toolchain 2019.03-rc1) 8.3.1 20190225 39 fedora:31 : Ok gcc (GCC) 9.2.1 20190827 (Red Hat 9.2.1-1), clang version 9.0.0 (Fedora 9.0.0-1.fc31) 40 fedora:32 : Ok gcc (GCC) 9.2.1 20190827 (Red Hat 9.2.1-1), clang version 9.0.0 (Fedora 9.0.0-1.fc32) 41 fedora:rawhide : Ok gcc (GCC) 9.2.1 20190827 (Red Hat 9.2.1-1), clang version 9.0.0 (Fedora 9.0.0-1.fc32) 42 gentoo-stage3-amd64:latest : Ok gcc (Gentoo 8.3.0-r1 p1.1) 8.3.0 43 mageia:5 : Ok gcc (GCC) 4.9.2, clang version 3.5.2 (tags/RELEASE_352/final) 44 mageia:6 : Ok gcc (Mageia 5.5.0-1.mga6) 5.5.0, clang version 3.9.1 (tags/RELEASE_391/final) 45 mageia:7 : Ok gcc (Mageia 8.3.1-0.20190524.1.mga7) 8.3.1 20190524, clang version 8.0.0 (Mageia 8.0.0-1.mga7) 46 manjaro:latest : Ok gcc (GCC) 9.2.0, clang version 8.0.1 (tags/RELEASE_801/final) 47 opensuse:15.0 : Ok gcc (SUSE Linux) 7.4.1 20190424 [gcc-7-branch revision 270538], clang version 5.0.1 (tags/RELEASE_501/final 312548) 48 opensuse:15.1 : Ok gcc (SUSE Linux) 7.4.1 20190424 [gcc-7-branch revision 270538], clang version 7.0.1 (tags/RELEASE_701/final 349238) 49 opensuse:42.3 : Ok gcc (SUSE Linux) 4.8.5, clang version 3.8.0 (tags/RELEASE_380/final 262553) 50 opensuse:tumbleweed : Ok gcc (SUSE Linux) 9.2.1 20190903 [gcc-9-branch revision 275330], clang version 8.0.1 (tags/RELEASE_801/final 366581) 51 oraclelinux:6 : Ok gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-23.0.1) 52 oraclelinux:7 : Ok gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-39.0.1), clang version 3.4.2 (tags/RELEASE_34/dot2-final) 53 oraclelinux:8 : Ok gcc (GCC) 8.2.1 20180905 (Red Hat 8.2.1-3.0.1), clang version 7.0.1 (tags/RELEASE_701/final) 54 ubuntu:12.04 : Ok gcc (Ubuntu/Linaro 4.6.3-1ubuntu5) 4.6.3, Ubuntu clang version 3.0-6ubuntu3 (tags/RELEASE_30/final) (based on LLVM 3.0) 55 ubuntu:14.04 : Ok gcc (Ubuntu 4.8.4-2ubuntu1~14.04.4) 4.8.4, Ubuntu clang version 3.4-1ubuntu3 (tags/RELEASE_34/final) (based on LLVM 3.4) 56 ubuntu:16.04 : Ok gcc (Ubuntu 5.4.0-6ubuntu1~16.04.11) 5.4.0 20160609, clang version 3.8.0-2ubuntu4 (tags/RELEASE_380/final) 57 ubuntu:16.04-x-arm : Ok arm-linux-gnueabihf-gcc (Ubuntu/Linaro 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609 58 ubuntu:16.04-x-arm64 : Ok aarch64-linux-gnu-gcc (Ubuntu/Linaro 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609 59 ubuntu:16.04-x-powerpc : Ok powerpc-linux-gnu-gcc (Ubuntu 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609 60 ubuntu:16.04-x-powerpc64 : Ok powerpc64-linux-gnu-gcc (Ubuntu/IBM 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609 61 ubuntu:16.04-x-powerpc64el : Ok powerpc64le-linux-gnu-gcc (Ubuntu/IBM 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609 62 ubuntu:16.04-x-s390 : Ok s390x-linux-gnu-gcc (Ubuntu 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609 63 ubuntu:18.04 : Ok gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0, clang version 6.0.0-1ubuntu2 (tags/RELEASE_600/final) 64 ubuntu:18.04-x-arm : Ok arm-linux-gnueabihf-gcc (Ubuntu/Linaro 7.4.0-1ubuntu1~18.04.1) 7.4.0 65 ubuntu:18.04-x-arm64 : Ok aarch64-linux-gnu-gcc (Ubuntu/Linaro 7.4.0-1ubuntu1~18.04.1) 7.4.0 66 ubuntu:18.04-x-m68k : Ok m68k-linux-gnu-gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0 67 ubuntu:18.04-x-powerpc : Ok powerpc-linux-gnu-gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0 68 ubuntu:18.04-x-powerpc64 : Ok powerpc64-linux-gnu-gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0 69 ubuntu:18.04-x-powerpc64el : Ok powerpc64le-linux-gnu-gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0 70 ubuntu:18.04-x-riscv64 : Ok riscv64-linux-gnu-gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0 71 ubuntu:18.04-x-s390 : Ok s390x-linux-gnu-gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0 72 ubuntu:18.04-x-sh4 : Ok sh4-linux-gnu-gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0 73 ubuntu:18.04-x-sparc64 : Ok sparc64-linux-gnu-gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0 74 ubuntu:18.10 : Ok gcc (Ubuntu 8.3.0-6ubuntu1~18.10.1) 8.3.0, clang version 7.0.0-3 (tags/RELEASE_700/final) 75 ubuntu:19.04 : Ok gcc (Ubuntu 8.3.0-6ubuntu1) 8.3.0, clang version 8.0.0-3 (tags/RELEASE_800/final) 76 ubuntu:19.04-x-alpha : Ok alpha-linux-gnu-gcc (Ubuntu 8.3.0-6ubuntu1) 8.3.0 77 ubuntu:19.04-x-arm64 : Ok aarch64-linux-gnu-gcc (Ubuntu/Linaro 8.3.0-6ubuntu1) 8.3.0 78 ubuntu:19.04-x-hppa : Ok hppa-linux-gnu-gcc (Ubuntu 8.3.0-6ubuntu1) 8.3.0 79 ubuntu:19.10 : Ok gcc (Ubuntu 9.2.1-8ubuntu1) 9.2.1 20190909, clang version 9.0.0-+rc5-1~exp1 (tags/RELEASE_900/rc5) # # uname -a Linux quaco 5.2.18-200.fc30.x86_64 #1 SMP Tue Oct 1 13:14:07 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux # git log --oneline -1 27198a893ba0 perf trace: Use STUL_STRARRAY_FLAGS with mmap # perf version --build-options perf version 5.4.rc3.g27198a893ba0 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 aio: [ on ] # HAVE_AIO_SUPPORT zstd: [ on ] # HAVE_ZSTD_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: time utils : Ok 60: map_groups__merge_in : Ok 61: x86 rdpmc : Ok 62: Convert perf time to TSC : Ok 63: DWARF unwind : Ok 64: x86 instruction decoder - new instructions : Ok 65: Intel PT packet decoder : Ok 66: x86 bp modify : Ok 67: probe libc's inet_pton & backtrace it with ping : Ok 68: Use vfs_getname probe to get syscall args filenames : Ok 69: Add vfs_getname probe to get syscall args filenames : Ok 70: Check open filename arg using perf trace + vfs_getname: Ok 71: Zstd perf.data compression/decompression : Ok
$ make -C tools/perf build-test
make: Entering directory '/home/acme/git/perf/tools/perf'
- tarpkg: ./tests/perf-targz-src-pkg .
make_with_babeltrace_O: make LIBBABELTRACE=1
make_perf_o_O: make perf.o
make_install_prefix_slash_O: make install prefix=/tmp/krava/
make_no_libnuma_O: make NO_LIBNUMA=1
make_no_libperl_O: make NO_LIBPERL=1
make_no_auxtrace_O: make NO_AUXTRACE=1
make_install_prefix_O: make install prefix=/tmp/krava
make_no_libaudit_O: make NO_LIBAUDIT=1
make_no_libunwind_O: make NO_LIBUNWIND=1
make_util_map_o_O: make util/map.o
make_no_gtk2_O: make NO_GTK2=1
make_no_scripts_O: make NO_LIBPYTHON=1 NO_LIBPERL=1
make_pure_O: make
make_no_libbpf_O: make NO_LIBBPF=1
make_clean_all_O: make clean all
make_install_bin_O: make install-bin
make_no_demangle_O: make NO_DEMANGLE=1
make_no_libpython_O: make NO_LIBPYTHON=1
make_debug_O: make DEBUG=1
make_no_newt_O: make NO_NEWT=1
make_no_libbionic_O: make NO_LIBBIONIC=1
make_tags_O: make tags
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 NO_LIBZSTD=1 NO_LIBCAP=1
make_doc_O: make doc
make_no_backtrace_O: make NO_BACKTRACE=1
make_util_pmu_bison_o_O: make util/pmu-bison.o
make_no_slang_O: make NO_SLANG=1
make_no_ui_O: make NO_NEWT=1 NO_SLANG=1 NO_GTK2=1
make_help_O: make help
make_static_O: make LDFLAGS=-static NO_PERF_READ_VDSO32=1 NO_PERF_READ_VDSOX32=1 NO_JVMTI=1
make_no_libelf_O: make NO_LIBELF=1
make_cscope_O: make cscope
make_no_libdw_dwarf_unwind_O: make NO_LIBDW_DWARF_UNWIND=1
make_install_O: make install
make_with_clangllvm_O: make LIBCLANGLLVM=1
OK
make: Leaving directory '/home/acme/git/perf/tools/perf'
$Arnaldo Carvalho de Melo
maintainerPATCH 1/57 perf tools: Allow to build with -ltcmalloc
From: Jiri Olsa <[email protected]>
By using "make TCMALLOC=1" you can enable perf to be build for usage with libtcmalloc.so (gperftools).
Get heap profile (tools/perf directory):
$ <install gperftools>
$ make TCMALLOC=1 DEBUG=1
$ HEAPPROFILE=/tmp/heapprof ./perf ...
$ pprof ./perf /tmp/heapprof.000*
(pprof) top
Total: 2335.5 MB
1735.1 74.3% 74.3% 1735.1 74.3% memdup
402.0 17.2% 91.5% 402.0 17.2% zalloc
140.2 6.0% 97.5% 145.8 6.2% map__new
33.6 1.4% 98.9% 33.6 1.4% symbol__new
12.4 0.5% 99.5% 12.4 0.5% alloc_event
6.2 0.3% 99.7% 6.2 0.3% nsinfo__new
5.5 0.2% 100.0% 5.5 0.2% nsinfo__copy
0.3 0.0% 100.0% 0.3 0.0% dso__new
0.1 0.0% 100.0% 0.1 0.0% do_read_string
0.0 0.0% 100.0% 0.0 0.0% __GI__IO_file_doallocateSee callstack:
$ pprof --pdf ./perf /tmp/heapprof.00* > callstack.pdf $ pprof --web ./perf /tmp/heapprof.00*
Committer testing:
Install gperftools, on fedora:
# dnf install gperftools-devel
Then build:
$ make TCMALLOC=1 DEBUG=1 -C tools/perf O=/tmp/build/perf install-bin
Verify that it linked against the right library:
$ ldd ~/bin/perf | grep tcma libtcmalloc.so.4 => /lib64/libtcmalloc.so.4 (0x00007fb2953a7000) $
Run 'perf trace' system wide for 1 minute:
# HEAPPROFILE=/tmp/heapprof perf trace -a sleep 1m
<SNIP>
59985.524 ( 0.006 ms): Web Content/20354 recvmsg(fd: 9<socket:[1762817]>, msg: 0x7ffee5fdafb0) = -1 EAGAIN (Resource temporarily unavailable) 59985.536 ( 0.005 ms): Web Content/20354 recvmsg(fd: 9<socket:[1762817]>, msg: 0x7ffee5fdafc0) = -1 EAGAIN (Resource temporarily unavailable)
59981.956 (10.143 ms): SCTP timer/21716 ... [continued]: select()) = 0 (Timeout)
59985.549 ( ): Web Content/20354 poll(ufds: 0x7f1df38af180, nfds: 3, timeout_msecs: 4294967295) ...
0.926 (59999.481 ms): sleep/29764 ... [continued]: nanosleep()) = 0 59992.133 ( ): SCTP timer/21716 select(tvp: 0x7ff5bf7fee80) ... 60000.477 ( 0.009 ms): sleep/29764 close(fd: 1) = 0 60000.493 ( 0.005 ms): sleep/29764 close(fd: 2) = 0 60000.514 ( ): sleep/29764 exit_group() = ?
Dumping heap profile to /tmp/heapprof.0001.heap (Exiting, 3 MB in use)
[root@quaco ~]#
Install pprof:
# dnf install pprof
And run it:
# pprof ~/bin/perf /tmp/heapprof.0001.heap
Using local file /root/bin/perf.
Using local file /tmp/heapprof.0001.heap.
Welcome to pprof! For help, type 'help'.
(pprof) top
Total: 4.0 MB
1.7 42.0% 42.0% 2.2 54.1% map__new
0.9 23.3% 65.3% 0.9 23.3% zalloc
0.5 11.4% 76.7% 0.5 11.4% dso__new
0.2 5.6% 82.3% 0.3 8.5% trace__sys_enter
0.2 4.9% 87.2% 0.2 4.9% __GI___strdup
0.2 3.8% 91.0% 0.2 3.8% new_term
0.1 2.2% 93.2% 0.4 10.1% __perf_pmu__new_alias
0.0 1.0% 94.3% 0.0 1.2% event_read_fields
0.0 0.8% 95.1% 0.0 0.8% nsinfo__new
0.0 0.7% 95.8% 0.1 3.2% trace__read_syscall_info
(pprof)--- tools/perf/Makefile.config | 5 +++++ tools/perf/Makefile.perf | 2 ++ 2 files changed, 7 insertions(+)
tools/perf/Makefile.config +5 -0
@@ -265,6 +265,11 @@ LDFLAGS += -Wl,-z,noexecstack EXTLIBS = -lpthread -lrt -lm -ldl +ifneq ($(TCMALLOC),) + CFLAGS += -fno-builtin-malloc -fno-builtin-calloc -fno-builtin-realloc -fno-builtin-free + EXTLIBS += -ltcmalloc +endif + ifeq ($(FEATURES_DUMP),) include $(srctree)/tools/build/Makefile.feature else
tools/perf/Makefile.perf +2 -0
@@ -114,6 +114,8 @@ include ../scripts/utilities.mak # Define NO_LIBZSTD if you do not want support of Zstandard based runtime # trace compression in record mode. # +# Define TCMALLOC to enable tcmalloc heap profiling. +# # As per kernel Makefile, avoid funny character set dependencies unexport LC_ALL
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 2/57 perf script: Fix --reltime with --time
In reply to Arnaldo Carvalho de MeloFrom: Andi Kleen <[email protected]>
My earlier patch to just enable --reltime with --time was a little too optimistic. The --time parsing would accept absolute time, which is very confusing to the user.
Support relative time in --time parsing too. This only works with recent perf record that records the first sample time. Otherwise we error out.
--- tools/perf/builtin-script.c | 5 +++-- tools/perf/util/time-utils.c | 27 ++++++++++++++++++++++++--- tools/perf/util/time-utils.h | 5 +++++ 3 files changed, 32 insertions(+), 5 deletions(-)
tools/perf/builtin-script.c +3 -2
@@ -3864,10 +3864,11 @@ int cmd_script(int argc, const char **argv) goto out_delete;if (script.time_str) { - err =perf_time__parse_for_ranges (script.time_str, session, + err =perf_time__parse_for_ranges_reltime (script.time_str, session, &script.ptime_range, &script.range_size, - &script.range_num); + &script.range_num, + reltime);if (err <0 ) goto out_delete;
tools/perf/util/time-utils.c +24 -3
@@ -458,10 +458,11 @@ bool perf_time__ranges_skip_sample(struct perf_time_interval *ptime_buf,return true; } -int perf_time__parse_for_ranges (const char *time_str, +int perf_time__parse_for_ranges_reltime (const char *time_str,struct perf_session *session,struct perf_time_interval **ranges, -int *range_size,int *range_num) +int *range_size,int *range_num, +bool reltime) {bool has_percent =strchr (time_str,'%' );struct perf_time_interval *ptime_range;
@@ -471,7 +472,7 @@ int perf_time__parse_for_ranges(const char *time_str,if (!ptime_range)return -ENOMEM ; -if (has_percent) { +if (has_percent || reltime) {if (session->evlist->first_sample_time ==0 && session->evlist->last_sample_time ==0 ) { pr_err("HINT: no first/last sample time found in perf data.\n"
@@ -479,7 +480,9 @@ int perf_time__parse_for_ranges(const char *time_str, "(if '--buildid-all' is enabled, please set '--timestamp-boundary').\n"); goto error; } + } +if (has_percent) { num = perf_time__percent_parse_str( ptime_range, size, time_str,
@@ -492,6 +495,15 @@ int perf_time__parse_for_ranges(const char *time_str,if (num <0 ) goto error_invalid; +if (reltime) { +int i; + +for (i =0 ; i < num; i++) { + ptime_range[i].start += session->evlist->first_sample_time; + ptime_range[i].end += session->evlist->first_sample_time; + } + } + *range_size = size; *range_num = num; *ranges = ptime_range;
@@ -504,6 +516,15 @@ int perf_time__parse_for_ranges(const char *time_str,return ret; } +int perf_time__parse_for_ranges (const char *time_str, +struct perf_session *session, +struct perf_time_interval **ranges, +int *range_size,int *range_num) +{ +return perf_time__parse_for_ranges_reltime (time_str, session, ranges, + range_size, range_num, false); +} +int timestamp__scnprintf_usec (u64 timestamp,char *buf,size_t sz) {u64 sec = timestamp /NSEC_PER_SEC ;
tools/perf/util/time-utils.h +5 -0
@@ -26,6 +26,11 @@ bool perf_time__ranges_skip_sample(struct perf_time_interval *ptime_buf,struct perf_session ; +int perf_time__parse_for_ranges_reltime (const char *str,struct perf_session *session, +struct perf_time_interval **ranges, +int *range_size,int *range_num, +bool reltime); +int perf_time__parse_for_ranges (const char *str,struct perf_session *session,struct perf_time_interval **ranges,int *range_size,int *range_num);
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 3/57 perf evlist: Fix fix for freed id arrays
In reply to Arnaldo Carvalho de MeloFrom: Andi Kleen <[email protected]>
In the earlier fix for the memory overrun of id arrays I managed to typo the wrong event in the fix.
Of course we need to close the current event in the loop, not the original failing event.
The same test case as in the original patch still passes.
--- tools/perf/util/evlist.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/util/evlist.c +1 -1
@@ -1599,7 +1599,7 @@ struct evsel *perf_evlist__reset_weak_group(struct evlist *evsel_list, is_open = false;if (c2->leader == leader) {if (is_open) -perf_evsel__close (&evsel->core); +perf_evsel__close (&c2->core); c2->leader = c2; c2->core.nr_members =0 ; }
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 4/57 perf test: Report failure for mmap events
In reply to Arnaldo Carvalho de MeloFrom: Leo Yan <[email protected]>
When fail to mmap events in task exit case, it misses to set 'err' to -1; thus the testing will not report failure for it.
This patch sets 'err' to -1 when fails to mmap events, thus Perf tool can report correct result.
--- tools/perf/tests/task-exit.c | 1 + 1 file changed, 1 insertion(+)
tools/perf/tests/task-exit.c +1 -0
@@ -111,6 +111,7 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unusedif (evlist__mmap (evlist,128 ) <0 ) {pr_debug ("failed to mmap events: %d (%s)\n" , errno,str_error_r (errno, sbuf,sizeof (sbuf))); + err =-1 ; goto out_delete_evlist; }
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 5/57 perf test: Avoid infinite loop for task exit case
In reply to Arnaldo Carvalho de MeloFrom: Leo Yan <[email protected]>
When executing the task exit testing case, perf gets stuck in an endless loop this case and doesn't return back on Arm64 Juno board.
After digging into this issue, since Juno board has Arm's big.LITTLE CPUs, thus the PMUs are not compatible between the big CPUs and little CPUs. This leads to a PMU event that cannot be enabled properly when the traced task is migrated from one variant's CPU to another variant. Finally, the test case runs into infinite loop for cannot read out any event data after return from polling.
Eventually, we need to work out formal solution to allow PMU events can be freely migrated from one CPU variant to another, but this is a difficult task and a different topic. This patch tries to fix the Perf test case to avoid infinite loop, when the testing detects 1000 times retrying for reading empty events, it will directly bail out and return failure. This allows the Perf tool can continue its other test cases.
--- tools/perf/tests/task-exit.c | 8 ++++++++ 1 file changed, 8 insertions(+)
tools/perf/tests/task-exit.c +8 -0
@@ -54,6 +54,7 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unusedstruct perf_cpu_map *cpus;struct perf_thread_map *threads;struct mmap *md; +int retry_count =0 ;signal (SIGCHLD , sig_handler);
@@ -133,6 +134,13 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unused out_init:if (!exited || !nr_exit) {evlist__poll (evlist,-1 ); + +if (retry_count++ >1000 ) { +pr_debug ("Failed after retrying 1000 times\n" ); + err =-1 ; + goto out_free_maps; + } + goto retry; }
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 6/57 perf report: Add warning when libunwind not compiled in
In reply to Arnaldo Carvalho de MeloFrom: Jin Yao <[email protected]>
We received a user report that call-graph DWARF mode was enabled in 'perf record' but 'perf report' didn't unwind the callstack correctly. The reason was, libunwind was not compiled in.
We can use 'perf -vv' to check the compiled libraries but it would be valuable to report a warning to user directly (especially valuable for a perf newbie).
The warning is:
Warning:
Please install libunwind development packages during the perf build.
Both TUI and stdio are supported.
--- tools/perf/builtin-report.c | 7 +++++++ 1 file changed, 7 insertions(+)
tools/perf/builtin-report.c +7 -0
@@ -399,6 +399,13 @@ static int report__setup_sample_type(struct report *rep) PERF_SAMPLE_BRANCH_ANY)) rep->nonany_branch_mode = true; +#ifndef HAVE_LIBUNWIND_SUPPORT +if (dwarf_callchain_users) { +ui__warning ("Please install libunwind development packages " +"during the perf build.\n" ); + } +#endif +return 0 ; }
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 7/57 perf annotate: Avoid reallocation in objdump parsing
In reply to Arnaldo Carvalho de MeloFrom: Ian Rogers <[email protected]>
Objdump output is parsed using getline which allocates memory for the read. Getline will realloc if the memory is too small, but currently the line is always freed after the call.
Simplify parse_objdump_line by performing the reading in symbol__disassemble.
--- tools/perf/util/annotate.c | 26 ++++++++++++++------------ 1 file changed, 14 insertions(+), 12 deletions(-)
tools/perf/util/annotate.c +14 -12
@@ -1489,24 +1489,17 @@ annotation_line__print(struct annotation_line *al, struct symbol *sym, u64 start* means that it's not a disassembly line so should be treated differently. * The ops.raw part will be parsed further according to type of the instruction. */ -static int symbol__parse_objdump_line (struct symbol *sym,FILE *file, +static int symbol__parse_objdump_line (struct symbol *sym,struct annotate_args *args, -int *line_nr) +char *line,int *line_nr) {struct map *map = args->ms.map;struct annotation *notes =symbol__annotation (sym);struct disasm_line *dl; -char *line =NULL , *parsed_line, *tmp, *tmp2; -size_t line_len; +char *parsed_line, *tmp, *tmp2;s64 line_ip, offset =-1 ;regmatch_t match[2 ]; -if (getline (&line, &line_len, file) <0 ) -return -1 ; - -if (!line) -return -1 ; - line_ip =-1 ; parsed_line =strim (line);
@@ -1543,7 +1536,6 @@ static int symbol__parse_objdump_line(struct symbol *sym, FILE *file, args->ms.sym = sym; dl =disasm_line__new (args); -free (line); (*line_nr)++;if (dl ==NULL )
@@ -1876,6 +1868,8 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args)int lineno =0 ;int nline;pid_t pid; +char *line; +size_t line_len;int err =dso__disassemble_filename (dso, symfs_filename,sizeof (symfs_filename));if (err)
@@ -1964,18 +1958,26 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args) goto out_free_command; } +/* Storage for getline. */ + line =NULL ; + line_len =0 ; + nline =0 ;while (!feof (file)) { +if (getline (&line, &line_len, file) <0 || !line) +break ; +/* * The source code line number (lineno) needs to be kept in * across calls to symbol__parse_objdump_line(), so that it * can associate it with the instructions till the next one. * See disasm_line__new() and struct disasm_line::line_nr. */ -if (symbol__parse_objdump_line (sym, file, args, &lineno) <0 ) +if (symbol__parse_objdump_line (sym, args, line, &lineno) <0 )break ; nline++; } +free (line);if (nline ==0 )pr_err ("No output from %s\n" , command);
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 8/57 perf annotate: Use libsubcmd's run-command.h to fork objdump
In reply to Arnaldo Carvalho de MeloFrom: Ian Rogers <[email protected]>
Reduce duplicated logic by using the subcmd library. Ensure when errors occur they are reported to the caller. Before this patch, if no lines are read the error status is 0.
--- tools/perf/util/annotate.c | 72 ++++++++++++++++++++------------------ 1 file changed, 37 insertions(+), 35 deletions(-)
tools/perf/util/annotate.c +37 -35
@@ -43,6 +43,7 @@#include <linux/string.h> #include <bpf/libbpf.h> #include <subcmd/parse-options.h> +#include <subcmd/run-command.h> /* FIXME: For the HE_COLORSET */ #include "ui/browser.h"
@@ -1864,12 +1865,19 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args)struct kcore_extract kce;bool delete_extract = false;bool decomp = false; -int stdout_fd[2 ];int lineno =0 ;int nline; -pid_t pid;char *line;size_t line_len; +const char *objdump_argv[] = { +"/bin/sh" , +"-c" , +NULL ,/* Will be the objdump command to run. */ +"--" , +NULL ,/* Will be the symfs path. */ +NULL , + }; +struct child_process objdump_process;int err =dso__disassemble_filename (dso, symfs_filename,sizeof (symfs_filename));if (err)
@@ -1899,7 +1907,7 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args)if (dso__decompress_kmodule_path (dso, symfs_filename, tmp,sizeof (tmp)) <0 ) - goto out; +return -1 ; decomp = true;strcpy (symfs_filename, tmp);
@@ -1924,38 +1932,28 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args)pr_debug ("Executing: %s\n" , command); - err =-1 ; -if (pipe (stdout_fd) <0 ) { -pr_err ("Failure creating the pipe to run %s\n" , command); - goto out_free_command; - } - - pid =fork (); -if (pid <0 ) { -pr_err ("Failure forking to run %s\n" , command); - goto out_close_stdout; - } + objdump_argv[2 ] = command; + objdump_argv[4 ] = symfs_filename; -if (pid ==0 ) { -close (stdout_fd[0 ]); -dup2 (stdout_fd[1 ],1 ); -close (stdout_fd[1 ]); -execl ("/bin/sh" ,"sh" ,"-c" , command,"--" , symfs_filename, -NULL ); -perror (command); -exit (-1 ); +/* Create a pipe to read from for stdout */ +memset (&objdump_process,0 ,sizeof (objdump_process)); + objdump_process.argv = objdump_argv; + objdump_process.out =-1 ; +if (start_command (&objdump_process)) { +pr_err ("Failure starting to run %s\n" , command); + err =-1 ; + goto out_free_command; } -close (stdout_fd[1 ]); - - file =fdopen (stdout_fd[0 ],"r" ); + file =fdopen (objdump_process.out,"r" );if (!file) {pr_err ("Failure creating FILE stream for %s\n" , command);/* * If we were using debug info should retry with * original binary. */ - goto out_free_command; + err =-1 ; + goto out_close_stdout; }/* Storage for getline. */
@@ -1979,8 +1977,14 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args) }free (line); -if (nline ==0 ) + err =finish_command (&objdump_process); +if (err) +pr_err ("Error running %s\n" , command); + +if (nline ==0 ) { + err =-1 ;pr_err ("No output from %s\n" , command); + }/* * kallsyms does not have symbol sizes so there may a nop at the end.
@@ -1990,23 +1994,21 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args)delete_last_nop (sym);fclose (file); - err =0 ; + +out_close_stdout: +close (objdump_process.out); + out_free_command:free (command); -out_remove_tmp: -close (stdout_fd[0 ]); +out_remove_tmp:if (decomp)unlink (symfs_filename);if (delete_extract)kcore_extract__delete (&kce); -out: -return err; -out_close_stdout: -close (stdout_fd[1 ]); - goto out_free_command; +return err; } static void calc_percent(struct sym_hist *sym_hist,
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 9/57 perf annotate: Don't pipe objdump output through 'grep' command
In reply to Arnaldo Carvalho de MeloFrom: Ian Rogers <[email protected]>
Simplify the objdump command by not piping the output of objdump through grep. Instead, drop lines that match the grep pattern during the reading loop.
--- tools/perf/util/annotate.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)
tools/perf/util/annotate.c +8 -1
@@ -1916,7 +1916,7 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args) err = asprintf(&command, "%s %s%s --start-address=0x%016" PRIx64 " --stop-address=0x%016" PRIx64 - " -l -d %s %s -C \"$1\" 2>/dev/null|grep -v \"$1:\"|expand", + " -l -d %s %s -C \"$1\" 2>/dev/null|expand", opts->objdump_path ?: "objdump", opts->disassembler_style ? "-M " : "", opts->disassembler_style ?: "",
@@ -1962,9 +1962,16 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args) nline =0 ;while (!feof (file)) { +const char *match; +if (getline (&line, &line_len, file) <0 || !line)break ; +/* Skip lines containing "filename:" */ + match =strstr (line, symfs_filename); +if (match && match[strlen (symfs_filename)] ==':' ) +continue ; +/* * The source code line number (lineno) needs to be kept in * across calls to symbol__parse_objdump_line(), so that it
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 10/57 perf annotate: Don't pipe objdump output through 'expand' command
In reply to Arnaldo Carvalho de MeloFrom: Ian Rogers <[email protected]>
Avoiding a pipe allows objdump command failures to surface. Move to the caller of symbol__parse_objdump_line the call to strim that removes leading and trailing tabs. Add a new expand_tabs function that if a tab is present allocate a new line in which tabs are expanded. In symbol__parse_objdump_line the line had no leading spaces, so simplify the line_ip processing.
--- tools/perf/util/annotate.c | 95 ++++++++++++++++++++++++++++++-------- 1 file changed, 76 insertions(+), 19 deletions(-)
tools/perf/util/annotate.c +76 -19
@@ -1492,35 +1492,24 @@ annotation_line__print(struct annotation_line *al, struct symbol *sym, u64 start*/ static int symbol__parse_objdump_line (struct symbol *sym,struct annotate_args *args, -char *line,int *line_nr) +char *parsed_line,int *line_nr) {struct map *map = args->ms.map;struct annotation *notes =symbol__annotation (sym);struct disasm_line *dl; -char *parsed_line, *tmp, *tmp2; +char *tmp;s64 line_ip, offset =-1 ;regmatch_t match[2 ]; - line_ip =-1 ; - parsed_line =strim (line); -/* /filename:linenr ? Save line number and ignore. */ if (regexec (&file_lineno, parsed_line,2 , match,0 ) ==0 ) { *line_nr =atoi (parsed_line + match[1 ].rm_so);return 0 ; } - tmp =skip_spaces (parsed_line); -if (*tmp) { -/* -* Parse hexa addresses followed by ':' -*/ - line_ip =strtoull (tmp, &tmp2,16 ); -if (*tmp2 !=':' || tmp == tmp2 || tmp2[1 ] =='\0' ) - line_ip =-1 ; - } - -if (line_ip !=-1 ) { +/* Process hex address followed by ':'. */ + line_ip =strtoull (parsed_line, &tmp,16 ); +if (parsed_line != tmp && tmp[0 ] ==':' && tmp[1 ] !='\0' ) {u64 start =map__rip_2objdump (map, sym->start), end =map__rip_2objdump (map, sym->end);
@@ -1528,7 +1517,7 @@ static int symbol__parse_objdump_line(struct symbol *sym,if ((u64 )line_ip < start || (u64 )line_ip >= end) offset =-1 ;else - parsed_line = tmp2 +1 ; + parsed_line = tmp +1 ; } args->offset = offset;
@@ -1854,6 +1843,67 @@ static int symbol__disassemble_bpf(struct symbol *sym __maybe_unused, }#endif // defined(HAVE_LIBBFD_SUPPORT) && defined(HAVE_LIBBPF_SUPPORT) +/* +* Possibly create a new version of line with tabs expanded. Returns the +* existing or new line, storage is updated if a new line is allocated. If +* allocation fails then NULL is returned. +*/ +static char *expand_tabs (char *line,char **storage,size_t *storage_len) +{ +size_t i, src, dst, len, new_storage_len, num_tabs; +char *new_line; +size_t line_len =strlen (line); + +for (num_tabs =0 , i =0 ; i < line_len; i++) +if (line[i] =='\t' ) + num_tabs++; + +if (num_tabs ==0 ) +return line; + +/* +* Space for the line and '\0', less the leading and trailing +* spaces. Each tab may introduce 7 additional spaces. +*/ + new_storage_len = line_len +1 + (num_tabs *7 ); + + new_line =malloc (new_storage_len); +if (new_line ==NULL ) { +pr_err ("Failure allocating memory for tab expansion\n" ); +return NULL ; + } + +/* +* Copy regions starting at src and expand tabs. If there are two +* adjacent tabs then 'src == i', the memcpy is of size 0 and the spaces +* are inserted. +*/ +for (i =0 , src =0 , dst =0 ; i < line_len && num_tabs; i++) { +if (line[i] =='\t' ) { + len = i - src; +memcpy (&new_line[dst], &line[src], len); + dst += len; + new_line[dst++] =' ' ; +while (dst %8 !=0 ) + new_line[dst++] =' ' ; + src = i +1 ; + num_tabs--; + } + } + +/* Expand the last region. */ + len = line_len +1 - src; +memcpy (&new_line[dst], &line[src], len); + dst += len; + new_line[dst] ='\0' ; + +free (*storage); + *storage = new_line; + *storage_len = new_storage_len; +return new_line; + +} +static int symbol__disassemble (struct symbol *sym,struct annotate_args *args) {struct annotation_options *opts = args->options;
@@ -1916,7 +1966,7 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args) err = asprintf(&command, "%s %s%s --start-address=0x%016" PRIx64 " --stop-address=0x%016" PRIx64 - " -l -d %s %s -C \"$1\" 2>/dev/null|expand", + " -l -d %s %s -C \"$1\" 2>/dev/null", opts->objdump_path ?: "objdump", opts->disassembler_style ? "-M " : "", opts->disassembler_style ?: "",
@@ -1963,6 +2013,7 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args) nline =0 ;while (!feof (file)) {const char *match; +char *expanded_line;if (getline (&line, &line_len, file) <0 || !line)break ;
@@ -1972,13 +2023,19 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args)if (match && match[strlen (symfs_filename)] ==':' )continue ; + expanded_line =strim (line); + expanded_line =expand_tabs (expanded_line, &line, &line_len); +if (!expanded_line) +break ; +/* * The source code line number (lineno) needs to be kept in * across calls to symbol__parse_objdump_line(), so that it * can associate it with the instructions till the next one. * See disasm_line__new() and struct disasm_line::line_nr. */ -if (symbol__parse_objdump_line (sym, args, line, &lineno) <0 ) +if (symbol__parse_objdump_line (sym, args, expanded_line, + &lineno) <0 )break ; nline++; }
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 11/57 perf annotate: Fix objdump --no-show-raw-insn flag
In reply to Arnaldo Carvalho de MeloFrom: Ian Rogers <[email protected]>
Remove redirection of objdump's stderr to /dev/null to help diagnose failures.
Fix the '--no-show-raw' flag to be '--no-show-raw-insn' which binutils is permissive and allows, but fails with LLVM objdump.
--- tools/perf/util/annotate.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
tools/perf/util/annotate.c +2 -2
@@ -1966,13 +1966,13 @@ static int symbol__disassemble(struct symbol *sym, struct annotate_args *args) err =asprintf (&command,"%s %s%s --start-address=0x%016" PRIx64" --stop-address=0x%016" PRIx64 -" -l -d %s %s -C \"$1\" 2>/dev/null" , +" -l -d %s %s -C \"$1\"" , opts->objdump_path ?:"objdump" , opts->disassembler_style ?"-M " :"" , opts->disassembler_style ?:"" ,map__rip_2objdump (map, sym->start),map__rip_2objdump (map, sym->end), - opts->show_asm_raw ?"" :"--no-show-raw" , + opts->show_asm_raw ?"" :"--no-show-raw-insn" , opts->annotate_src ?"-S" :"" );if (err <0 ) {
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 12/57 perf jvmti: Link against tools/lib/ctype.h to have weak strlcpy()
In reply to Arnaldo Carvalho de MeloFrom: Thomas Richter <[email protected]>
The build of file libperf-jvmti.so succeeds but the resulting object fails to load:
# ~/linux/tools/perf/perf record -k mono -- java \
-XX:+PreserveFramePointer \
-agentpath:/root/linux/tools/perf/libperf-jvmti.so \
hog 100000 123450 Error occurred during initialization of VM
Could not find agent library /root/linux/tools/perf/libperf-jvmti.so
in absolute path, with error:
/root/linux/tools/perf/libperf-jvmti.so: undefined symbol: _ctype
Add the missing _ctype symbol into the build script.
--- tools/perf/jvmti/Build | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)
tools/perf/jvmti/Build +5 -1
@@ -2,7 +2,7 @@ jvmti-y += libjvmti.o jvmti-y += jvmti_agent.o # For strlcpy -jvmti-y += libstring.o +jvmti-y += libstring.o libctype.o CFLAGS_jvmti = -fPIC -DPIC -I$(JDIR)/include -I$(JDIR)/include/linux CFLAGS_REMOVE_jvmti = -Wmissing-declarations
@@ -15,3 +15,7 @@ CFLAGS_libstring.o += -Wno-unused-parameter -DETC_PERFCONFIG="BUILD_STR($(ETC_PE $(OUTPUT)jvmti/libstring.o: ../lib/string.c FORCE $(call rule_mkdir) $(call if_changed_dep,cc_o_c) + +$(OUTPUT)jvmti/libctype.o: ../lib/ctype.c FORCE + $(call rule_mkdir) + $(call if_changed_dep,cc_o_c)
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 13/57 perf stat: Support --all-kernel/--all-user
In reply to Arnaldo Carvalho de MeloFrom: Jin Yao <[email protected]>
'perf record' has supported --all-kernel / --all-user to configure all used events to run in kernel space or run in user space. But 'perf stat' doesn't support these options.
It would be useful to support these options in 'perf stat' too to keep the same semantics available in both tools.
--- tools/perf/Documentation/perf-stat.txt | 6 ++++++ tools/perf/builtin-stat.c | 6 ++++++ tools/perf/util/stat.c | 10 ++++++++++ tools/perf/util/stat.h | 2 ++ 4 files changed, 24 insertions(+)
tools/perf/Documentation/perf-stat.txt +6 -0
@@ -323,6 +323,12 @@ The output is SMI cycles%, equals to (aperf - unhalted core cycles) / aperf Users who wants to get the actual value can apply --no-metric-only. +--all-kernel:: +Configure all used events to run in kernel space. + +--all-user:: +Configure all used events to run in user space. + EXAMPLES --------
tools/perf/builtin-stat.c +6 -0
@@ -803,6 +803,12 @@ static struct option stat_options[] = {OPT_CALLBACK ('M' ,"metrics" , &evsel_list,"metric/metric group list" ,"monitor specified metrics or metric groups (separated by ,)" , parse_metric_groups), +OPT_BOOLEAN_FLAG (0 ,"all-kernel" , &stat_config.all_kernel, +"Configure all used events to run in kernel space." , +PARSE_OPT_EXCLUSIVE ), +OPT_BOOLEAN_FLAG (0 ,"all-user" , &stat_config.all_user, +"Configure all used events to run in user space." , +PARSE_OPT_EXCLUSIVE ), OPT_END() };
tools/perf/util/stat.c +10 -0
@@ -490,6 +490,16 @@ int create_perf_stat_counter(struct evsel *evsel,if (config->identifier) attr->sample_type =PERF_SAMPLE_IDENTIFIER ; +if (config->all_user) { + attr->exclude_kernel =1 ; + attr->exclude_user =0 ; + } + +if (config->all_kernel) { + attr->exclude_kernel =0 ; + attr->exclude_user =1 ; + } +/* * Disabling all counters initially, they will be enabled * either manually by us or by kernel via enable_on_exec
tools/perf/util/stat.h +2 -0
@@ -106,6 +106,8 @@ struct perf_stat_config {bool big_num;bool no_merge;bool walltime_run_table; +bool all_kernel; +bool all_user;FILE *output;unsigned int interval;unsigned int timeout;
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 14/57 perf trace: Add syscall failure stats to -s/--summary and -S/--with-summary
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
Just like strace has:
# trace -s sleep 1
Summary of events:
sleep (32370), 80 events, 93.0%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
--------------- -------- ------ -------- --------- --------- --------- ------
nanosleep 1 0 1000.402 1000.402 1000.402 1000.402 0.00%
mmap 8 0 0.023 0.002 0.003 0.004 8.49%
close 5 0 0.015 0.001 0.003 0.009 51.39%
mprotect 4 0 0.014 0.002 0.003 0.005 16.95%
openat 3 0 0.013 0.003 0.004 0.005 14.29%
munmap 1 0 0.010 0.010 0.010 0.010 0.00%
read 4 0 0.005 0.001 0.001 0.002 16.83%
brk 4 0 0.004 0.001 0.001 0.002 20.82%
access 1 1 0.004 0.004 0.004 0.004 0.00%
fstat 3 0 0.003 0.001 0.001 0.001 12.17%
lseek 3 0 0.003 0.001 0.001 0.001 11.45%
arch_prctl 2 1 0.002 0.001 0.001 0.001 2.30%
execve 1 0 0.000 0.000 0.000 0.000 0.00%#
# perf trace -S sleep 1
? ... [continued]: execve()) = 0
0.028 brk(brk: NULL) = 0x559f5bd96000
0.033 arch_prctl(option: 0x3001, arg2: 0x7ffda8b715a0) = -1 EINVAL (Invalid argument)
0.046 access(filename: "/etc/ld.so.preload", mode: R) = -1 ENOENT (No such file or directory)
0.055 openat(dfd: CWD, filename: "/etc/ld.so.cache", flags: RDONLY|CLOEXEC) = 3
0.060 fstat(fd: 3, statbuf: 0x7ffda8b707a0) = 0
0.062 mmap(addr: NULL, len: 134346, prot: READ, flags: PRIVATE, fd: 3, off: 0) = 0x7f3aedfc4000
0.066 close(fd: 3) = 0
0.079 openat(dfd: CWD, filename: "/lib64/libc.so.6", flags: RDONLY|CLOEXEC) = 3
0.085 read(fd: 3, buf: 0x7ffda8b70948, count: 832) = 832
0.088 lseek(fd: 3, offset: 792, whence: SET) = 792
0.090 read(fd: 3, buf: 0x7ffda8b70810, count: 68) = 68
0.093 fstat(fd: 3, statbuf: 0x7ffda8b707f0) = 0
0.095 mmap(addr: NULL, len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS) = 0x7f3aedfc2000
0.101 lseek(fd: 3, offset: 792, whence: SET) = 792
0.103 read(fd: 3, buf: 0x7ffda8b70450, count: 68) = 68
0.105 lseek(fd: 3, offset: 864, whence: SET) = 864
0.107 read(fd: 3, buf: 0x7ffda8b70470, count: 32) = 32
0.110 mmap(addr: NULL, len: 1857472, prot: READ, flags: PRIVATE|DENYWRITE, fd: 3, off: 0) = 0x7f3aeddfc000
0.114 mprotect(start: 0x7f3aede1e000, len: 1679360, prot: NONE) = 0
0.121 mmap(addr: 0x7f3aede1e000, len: 1363968, prot: READ|EXEC, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x22000) = 0x7f3aede1e000
0.127 mmap(addr: 0x7f3aedf6b000, len: 311296, prot: READ, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x16f000) = 0x7f3aedf6b000
0.131 mmap(addr: 0x7f3aedfb8000, len: 24576, prot: READ|WRITE, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x1bb000) = 0x7f3aedfb8000
0.138 mmap(addr: 0x7f3aedfbe000, len: 14272, prot: READ|WRITE, flags: PRIVATE|FIXED|ANONYMOUS) = 0x7f3aedfbe000
0.147 close(fd: 3) = 0
0.158 arch_prctl(option: SET_FS, arg2: 0x7f3aedfc3580) = 0
0.210 mprotect(start: 0x7f3aedfb8000, len: 16384, prot: READ) = 0
0.230 mprotect(start: 0x559f5b27d000, len: 4096, prot: READ) = 0
0.236 mprotect(start: 0x7f3aee00f000, len: 4096, prot: READ) = 0
0.240 munmap(addr: 0x7f3aedfc4000, len: 134346) = 0
0.300 brk(brk: NULL) = 0x559f5bd96000
0.302 brk(brk: 0x559f5bdb7000) = 0x559f5bdb7000
0.305 brk(brk: NULL) = 0x559f5bdb7000
0.310 openat(dfd: CWD, filename: "/usr/lib/locale/locale-archive", flags: RDONLY|CLOEXEC) = 3
0.315 fstat(fd: 3, statbuf: 0x7f3aedfbdac0) = 0
0.318 mmap(addr: NULL, len: 217750512, prot: READ, flags: PRIVATE, fd: 3, off: 0) = 0x7f3ae0e52000
0.325 close(fd: 3) = 0
0.358 nanosleep(rqtp: 0x7ffda8b714b0, rmtp: NULL) = 0
1000.622 close(fd: 1) = 0
1000.641 close(fd: 2) = 0
1000.664 exit_group(error_code: 0) = ?Summary of events:
sleep (722), 80 events, 93.0%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
--------------- -------- ------ -------- --------- --------- --------- ------
nanosleep 1 0 1000.194 1000.194 1000.194 1000.194 0.00%
mmap 8 0 0.025 0.002 0.003 0.005 10.17%
close 5 0 0.018 0.001 0.004 0.010 50.18%
mprotect 4 0 0.016 0.003 0.004 0.006 16.81%
openat 3 0 0.011 0.003 0.004 0.004 6.57%
munmap 1 0 0.010 0.010 0.010 0.010 0.00%
brk 4 0 0.005 0.001 0.001 0.002 20.72%
read 4 0 0.005 0.001 0.001 0.002 16.71%
access 1 1 0.005 0.005 0.005 0.005 0.00%
fstat 3 0 0.004 0.001 0.001 0.002 14.82%
lseek 3 0 0.003 0.001 0.001 0.001 11.66%
arch_prctl 2 1 0.002 0.001 0.001 0.001 3.59%
execve 1 0 0.000 0.000 0.000 0.000 0.00%#
Works for system wide, e.g. for 1ms:
# perf trace -s -a sleep 0.001
Summary of events:
sleep (768), 94 events, 37.9%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
--------------- -------- ------ -------- --------- --------- --------- ------
nanosleep 1 0 1.133 1.133 1.133 1.133 0.00%
execve 7 6 0.351 0.003 0.050 0.316 88.53%
mmap 8 0 0.024 0.002 0.003 0.004 8.86%
mprotect 4 0 0.017 0.003 0.004 0.006 16.02%
openat 3 0 0.013 0.004 0.004 0.005 8.34%
munmap 1 0 0.010 0.010 0.010 0.010 0.00%
brk 4 0 0.007 0.001 0.002 0.002 10.99%
close 5 0 0.005 0.001 0.001 0.002 11.69%
read 5 0 0.005 0.000 0.001 0.002 30.53%
access 1 1 0.004 0.004 0.004 0.004 0.00%
fstat 3 0 0.004 0.001 0.001 0.002 10.74%
lseek 3 0 0.003 0.001 0.001 0.001 10.20%
arch_prctl 2 1 0.002 0.001 0.001 0.001 3.34%Web Content (21258), 46 events, 18.5%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
--------------- -------- ------ -------- --------- --------- --------- ------
recvmsg 12 12 0.015 0.001 0.001 0.002 8.50%
futex 2 0 0.008 0.003 0.004 0.005 27.08%
poll 6 0 0.006 0.000 0.001 0.002 22.14%
read 2 0 0.006 0.002 0.003 0.003 26.08%
write 1 0 0.002 0.002 0.002 0.002 0.00%Web Content (4365), 36 events, 14.5%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
--------------- -------- ------ -------- --------- --------- --------- ------
recvmsg 10 10 0.015 0.001 0.002 0.003 11.83%
poll 5 0 0.006 0.000 0.001 0.002 28.44%
futex 2 0 0.005 0.001 0.003 0.004 48.29%
read 1 0 0.003 0.003 0.003 0.003 0.00%Timer (21275), 14 events, 5.6%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
--------------- -------- ------ -------- --------- --------- --------- ------
futex 6 1 0.240 0.000 0.040 0.149 64.58%
write 1 0 0.008 0.008 0.008 0.008 0.00%Timer (4383), 14 events, 5.6%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
--------------- -------- ------ -------- --------- --------- --------- ------
futex 6 2 0.186 0.000 0.031 0.181 96.45%
write 1 0 0.010 0.010 0.010 0.010 0.00%Web Content (20354), 28 events, 11.3%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
--------------- -------- ------ -------- --------- --------- --------- ------
recvmsg 8 8 0.010 0.001 0.001 0.002 15.24%
poll 4 0 0.004 0.000 0.001 0.002 35.68%
futex 1 0 0.003 0.003 0.003 0.003 0.00%
read 1 0 0.003 0.003 0.003 0.003 0.00%Timer (20371), 10 events, 4.0%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
--------------- -------- ------ -------- --------- --------- --------- ------
futex 4 1 0.077 0.000 0.019 0.075 95.46%
write 1 0 0.005 0.005 0.005 0.005 0.00%[root@quaco ~]#
--- tools/perf/builtin-trace.c | 58 ++++++++++++++++++++++---------------- 1 file changed, 34 insertions(+), 24 deletions(-)
tools/perf/builtin-trace.c +34 -24
@@ -1958,11 +1958,16 @@ static struct syscall *trace__syscall_info(struct trace *trace,return NULL ; } +struct syscall_stats { +struct stats stats; +u64 nr_failures; +}; +static void thread__update_stats (struct thread_trace *ttrace, -int id,struct perf_sample *sample) +int id,struct perf_sample *sample,long err) {struct int_node *inode; -struct stats *stats; +struct syscall_stats *stats;u64 duration =0 ; inode =intlist__findnew (ttrace->syscall_stats, id);
@@ -1971,17 +1976,22 @@ static void thread__update_stats(struct thread_trace *ttrace, stats = inode->priv;if (stats ==NULL ) { - stats =malloc (sizeof (struct stats )); + stats =malloc (sizeof (*stats));if (stats ==NULL )return ; -init_stats (stats); + + stats->nr_failures =0 ; +init_stats (&stats->stats); inode->priv = stats; }if (ttrace->entry_time && sample->time > ttrace->entry_time) duration = sample->time - ttrace->entry_time; -update_stats (stats, duration); +update_stats (&stats->stats, duration); + +if (err <0 ) + ++stats->nr_failures; }static int trace__printf_interrupted_entry (struct trace *trace)
@@ -2226,11 +2236,11 @@ static int trace__sys_exit(struct trace *trace, struct evsel *evsel,trace__fprintf_sample (trace, evsel, sample, thread); -if (trace->summary) -thread__update_stats (ttrace, id, sample); - ret =perf_evsel__sc_tp_uint (evsel, ret, sample); +if (trace->summary) +thread__update_stats (ttrace, id, sample, ret); +if (!trace->fd_path_disabled && sc->is_open && ret >=0 && ttrace->filename.pending_open) {trace__set_fd_pathname (thread, ret, ttrace->filename.name); ttrace->filename.pending_open = false;
@@ -4016,17 +4026,17 @@ static size_t trace__fprintf_threads_header(FILE *fp) }DEFINE_RESORT_RB (syscall_stats, a->msecs > b->msecs, - struct stats *stats; -double msecs; -int syscall; + struct syscall_stats *stats; +double msecs; +int syscall; ) { struct int_node *source = rb_entry(nd, struct int_node, rb_node); -struct stats *stats = source->priv; +struct syscall_stats *stats = source->priv; entry->syscall = source->i; entry->stats = stats; - entry->msecs = stats ? (u64 )stats->n * (avg_stats (stats) /NSEC_PER_MSEC ) :0 ; + entry->msecs = stats ? (u64 )stats->stats.n * (avg_stats (&stats->stats) /NSEC_PER_MSEC ) :0 ; } static size_t thread__dump_stats(struct thread_trace *ttrace,
@@ -4042,26 +4052,26 @@ static size_t thread__dump_stats(struct thread_trace *ttrace, printed +=fprintf (fp,"\n" ); - printed +=fprintf (fp," syscall calls total min avg max stddev\n" ); - printed +=fprintf (fp," (msec) (msec) (msec) (msec) (%%)\n" ); - printed +=fprintf (fp," --------------- -------- --------- --------- --------- --------- ------\n" ); + printed +=fprintf (fp," syscall calls errors total min avg max stddev\n" ); + printed +=fprintf (fp," (msec) (msec) (msec) (msec) (%%)\n" ); + printed +=fprintf (fp," --------------- -------- ------ -------- --------- --------- --------- ------\n" );resort_rb__for_each_entry (nd, syscall_stats) { -struct stats *stats = syscall_stats_entry->stats; +struct syscall_stats *stats = syscall_stats_entry->stats;if (stats) { -double min = (double )(stats->min) /NSEC_PER_MSEC ; -double max = (double )(stats->max) /NSEC_PER_MSEC ; -double avg =avg_stats (stats); +double min = (double )(stats->stats.min) /NSEC_PER_MSEC ; +double max = (double )(stats->stats.max) /NSEC_PER_MSEC ; +double avg =avg_stats (&stats->stats);double pct; -u64 n = (u64 ) stats->n; +u64 n = (u64 )stats->stats.n; - pct = avg ?100.0 *stddev_stats (stats)/avg :0.0 ; + pct = avg ?100.0 *stddev_stats (&stats->stats) / avg :0.0 ; avg /=NSEC_PER_MSEC ; sc = &trace->syscalls.table[syscall_stats_entry->syscall]; printed +=fprintf (fp," %-15s" , sc->name); - printed +=fprintf (fp," %8" PRIu64" %9.3f %9.3f %9.3f" , - n, syscall_stats_entry->msecs, min, avg); + printed +=fprintf (fp," %8" PRIu64" %6" PRIu64" %9.3f %9.3f %9.3f" , + n, stats->nr_failures, syscall_stats_entry->msecs, min, avg); printed +=fprintf (fp," %9.3f %9.2f%%\n" , max, pct); } }
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 15/57 perf trace: Introduce --errno-summary
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
To be used with -S or -s, using just this new option implies -s, examples:
# perf trace --errno-summary sleep 1
Summary of events:
sleep (10793), 80 events, 93.0%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
--------------- -------- ------ -------- --------- --------- --------- ------
nanosleep 1 0 1000.427 1000.427 1000.427 1000.427 0.00%
mmap 8 0 0.026 0.002 0.003 0.005 9.18%
close 5 0 0.018 0.001 0.004 0.009 48.97%
mprotect 4 0 0.017 0.003 0.004 0.006 16.49%
openat 3 0 0.012 0.003 0.004 0.005 9.41%
munmap 1 0 0.010 0.010 0.010 0.010 0.00%
brk 4 0 0.005 0.001 0.001 0.002 22.77%
read 4 0 0.005 0.001 0.001 0.002 22.33%
access 1 1 0.004 0.004 0.004 0.004 0.00%
ENOENT: 1
fstat 3 0 0.004 0.001 0.001 0.002 17.18%
lseek 3 0 0.003 0.001 0.001 0.001 11.62%
arch_prctl 2 1 0.002 0.001 0.001 0.001 3.32%
EINVAL: 1
execve 1 0 0.000 0.000 0.000 0.000 0.00%#
Works as well together with --failure and -S, i.e. collect the stats and show just the syscalls that failed:
# perf trace --failure -S --errno-summary sleep 1
0.032 arch_prctl(option: 0x3001, arg2: 0x7fffdb11b580) = -1 EINVAL (Invalid argument)
0.045 access(filename: "/etc/ld.so.preload", mode: R) = -1 ENOENT (No such file or directory)Summary of events:
sleep (10806), 80 events, 93.0%
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
--------------- -------- ------ -------- --------- --------- --------- ------
nanosleep 1 0 1000.094 1000.094 1000.094 1000.094 0.00%
mmap 8 0 0.026 0.002 0.003 0.005 9.06%
close 5 0 0.018 0.001 0.004 0.010 49.58%
mprotect 4 0 0.017 0.003 0.004 0.006 17.56%
openat 3 0 0.014 0.004 0.005 0.006 12.29%
munmap 1 0 0.010 0.010 0.010 0.010 0.00%
brk 4 0 0.005 0.001 0.001 0.002 22.75%
read 4 0 0.005 0.001 0.001 0.002 17.19%
access 1 1 0.005 0.005 0.005 0.005 0.00%
ENOENT: 1
fstat 3 0 0.004 0.001 0.001 0.002 21.66%
lseek 3 0 0.003 0.001 0.001 0.001 11.71%
arch_prctl 2 1 0.002 0.001 0.001 0.001 2.66%
EINVAL: 1
execve 1 0 0.000 0.000 0.000 0.000 0.00%#
--- tools/perf/Documentation/perf-trace.txt | 4 ++ tools/perf/builtin-trace.c | 51 +++++++++++++++++++++++-- 2 files changed, 51 insertions(+), 4 deletions(-)
tools/perf/Documentation/perf-trace.txt +4 -0
@@ -146,6 +146,10 @@ the thread executes on the designated CPUs. Default is to monitor all CPUs. Show all syscalls followed by a summary by thread with min, max, and average times (in msec) and relative stddev. +--errno-summary:: + To be used with -s or -S, to show stats for the errnos experienced by + syscalls, using only this option will trigger --summary. + --tool_stats:: Show tool stats such as number of times fd->pathname was discovered thru hooking the open syscall return + vfs_getname or via reading /proc/pid/fd, etc.
tools/perf/builtin-trace.c +47 -4
102 lines not on this page · raw message
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 16/57 perf string: Export asprintf__tp_filter_pids()
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
Will be used directly in 'perf trace' for setting up the command line argv array to pass to cmd_record, as this was how 'perf trace record' was implemented, following the model used in 'perf kvm record', 'perf sched record', etc.
--- tools/perf/util/evlist.c | 3 ++- tools/perf/util/string2.h | 3 +++ 2 files changed, 5 insertions(+), 1 deletion(-)
tools/perf/util/evlist.c +2 -1
@@ -21,6 +21,7 @@#include "../perf.h" #include "asm/bug.h" #include "bpf-event.h" +#include "util/string2.h" #include <signal.h> #include <unistd.h> #include <sched.h>
@@ -959,7 +960,7 @@ int perf_evlist__append_tp_filter(struct evlist *evlist, const char *filter)return err; } -static char *asprintf__tp_filter_pids (size_t npids,pid_t *pids) +char *asprintf__tp_filter_pids (size_t npids,pid_t *pids) {char *filter;size_t i;
tools/perf/util/string2.h +3 -0
@@ -4,6 +4,7 @@#include <linux/string.h> #include <linux/types.h> +#include <sys/types.h> // pid_t #include <stddef.h> #include <string.h>
@@ -32,6 +33,8 @@ static inline char *asprintf_expr_not_in_ints(const char *var, size_t nints, intreturn asprintf_expr_inout_ints (var, false, nints, ints); } +char *asprintf__tp_filter_pids (size_t npids,pid_t *pids); +char *strpbrk_esc (char *str,const char *stopset);char *strdup_esc (const char *str);
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 17/57 perf trace: Filter own pid to avoid a feedback look in 'perf trace record -a'
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
When doing a system wide 'perf trace record' we need, just like in 'perf trace' live mode, to filter out perf trace's own pid, so set up a tracepoint filter for the raw_syscalls tracepoints right after adding them to the argv array that is set up to then call cmd_record().
--- tools/perf/builtin-trace.c | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-)
tools/perf/builtin-trace.c +16 -8
@@ -2796,21 +2796,23 @@ static int trace__record(struct trace *trace, int argc, const char **argv)"-m" ,"1024" , "-c", "1", }; - +pid_t pid =getpid (); +char *filter =asprintf__tp_filter_pids (1 , &pid);const char *const sc_args[] = {"-e" , };unsigned int sc_args_nr =ARRAY_SIZE (sc_args);const char *const majpf_args[] = {"-e" ,"major-faults" };unsigned int majpf_args_nr =ARRAY_SIZE (majpf_args);const char *const minpf_args[] = {"-e" ,"minor-faults" };unsigned int minpf_args_nr =ARRAY_SIZE (minpf_args); +int err =-1 ; -/* +1 is for the event string below */ - rec_argc =ARRAY_SIZE (record_args) + sc_args_nr +1 + +/* +3 is for the event string below and the pid filter */ + rec_argc =ARRAY_SIZE (record_args) + sc_args_nr +3 + majpf_args_nr + minpf_args_nr + argc; rec_argv =calloc (rec_argc +1 ,sizeof (char *)); -if (rec_argv ==NULL ) -return -ENOMEM ; +if (rec_argv ==NULL || filter ==NULL ) + goto out_free; j =0 ;for (i =0 ; i <ARRAY_SIZE (record_args); i++)
@@ -2827,11 +2829,13 @@ static int trace__record(struct trace *trace, int argc, const char **argv) rec_argv[j++] ="syscalls:sys_enter,syscalls:sys_exit" ; else {pr_err ("Neither raw_syscalls nor syscalls events exist.\n" ); -free (rec_argv); -return -1 ; + goto out_free; } } + rec_argv[j++] ="--filter" ; + rec_argv[j++] = filter; +if (trace->trace_pgfaults &TRACE_PFMAJ )for (i =0 ; i < majpf_args_nr; i++) rec_argv[j++] = majpf_args[i];
@@ -2843,7 +2847,11 @@ static int trace__record(struct trace *trace, int argc, const char **argv)for (i =0 ; i < (unsigned int )argc; i++) rec_argv[j++] = argv[i]; -return cmd_record (j, rec_argv); + err =cmd_record (j, rec_argv); +out_free: +free (filter); +free (rec_argv); +return err; }static size_t trace__fprintf_thread_summary (struct trace *trace,FILE *fp);
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 18/57 perf trace: Support tracepoint dynamic char arrays
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
Things like:
# grep __data_loc /sys/kernel/debug/tracing/events/sched/sched_process_exec/format field:__data_loc char[] filename; offset:8; size:4; signed:1; #
That, at that offset (8) and with that size(8) have an integer that contains the real length and offset for the contents of that array.
Now this works:
# perf trace --max-events 1 -e sched:*exec -a
0.000 sed/19441 sched:sched_process_exec(filename: "/usr/bin/sync", pid: 19441 (sync), old_pid: 19441 (sync))
#As when using the libtraceevent based beautifier:
# perf trace --libtraceevent --max-events 1 -e sched:*exec -a
0.000 sync/19463 sched:sched_process_exec(filename=/usr/bin/sync pid=19463 old_pid=19463)
#I.e. that 'filename' is implemented as a dynamic char array.
--- tools/perf/builtin-trace.c | 19 ++++++++++++++----- tools/perf/trace/beauty/beauty.h | 2 ++ 2 files changed, 16 insertions(+), 5 deletions(-)
tools/perf/builtin-trace.c +14 -5
@@ -563,7 +563,7 @@ static size_t syscall_arg__scnprintf_char_array(char *bf, size_t size, struct sy// XXX Hey, maybe for sched:sched_switch prev/next comm fields we can // fill missing comms using thread__set_comm()... // here or in a special syscall_arg__scnprintf_pid_sched_tp... -return scnprintf (bf, size,"\"%-.*s\"" , arg->fmt->nr_entries, arg->val); +return scnprintf (bf, size,"\"%-.*s\"" , arg->fmt->nr_entries ?: arg->len, arg->val); }#define SCA_CHAR_ARRAY syscall_arg__scnprintf_char_array
@@ -1559,7 +1559,7 @@ syscall_arg_fmt__init_array(struct syscall_arg_fmt *arg, struct tep_format_field arg->scnprintf =SCA_PID ; else if (strcmp(field->type, "umode_t") == 0) arg->scnprintf = SCA_MODE_T; - else if ((field->flags & TEP_FIELD_IS_ARRAY) && strstarts(field->type, "char")) { + else if ((field->flags & TEP_FIELD_IS_ARRAY) && strstr(field->type, "char")) { arg->scnprintf =SCA_CHAR_ARRAY ; arg->nr_entries = field->arraylen; } else if ((strcmp(field->type, "int") == 0 ||
@@ -2523,10 +2523,19 @@ static size_t trace__fprintf_tp_fields(struct trace *trace, struct evsel *evsel,if (syscall_arg.mask & bit)continue ; + syscall_arg.len =0 ; syscall_arg.fmt = arg; -if (field->flags &TEP_FIELD_IS_ARRAY ) - val = (uintptr_t )(sample->raw_data + field->offset); -else +if (field->flags &TEP_FIELD_IS_ARRAY ) { +int offset = field->offset; + +if (field->flags &TEP_FIELD_IS_DYNAMIC ) { + offset =format_field__intval (field, sample, evsel->needs_swap); + syscall_arg.len = offset >>16 ; + offset &=0xffff ; + } + + val = (uintptr_t )(sample->raw_data + offset); + }else val =format_field__intval (field, sample, evsel->needs_swap);/* * Some syscall args need some mask, most don't and
tools/perf/trace/beauty/beauty.h +2 -0
@@ -87,6 +87,7 @@ struct syscall_arg_fmt;/** * @val: value of syscall argument being formatted +* @len: for tracepoint dynamic arrays, if fmt->nr_entries == 0, then its not a fixed array, look at arg->len * @args: All the args, use syscall_args__val(arg, nth) to access one * @augmented_args: Extra data that can be collected, for instance, with eBPF for expanding the pathname for open, etc * @augmented_args_size: augmented_args total payload size
@@ -109,6 +110,7 @@ struct syscall_arg {struct thread *thread;struct trace *trace;void *parm; +u16 len;u8 idx;u8 mask;bool show_string_prefix;
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 19/57 perf vendor events arm64: Fix Hisi hip08 DDRC PMU eventname
In reply to Arnaldo Carvalho de MeloFrom: John Garry <[email protected]>
The "EventName" for the DDRC precharge command event is incorrect, so fix it.
--- .../perf/pmu-events/arch/arm64/hisilicon/hip08/uncore-ddrc.json | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/pmu-events/arch/arm64/hisilicon/hip08/uncore-ddrc.json +1 -1
@@ -15,7 +15,7 @@ }, { "EventCode": "0x04", - "EventName": "uncore_hisi_ddrc.flux_wr", + "EventName": "uncore_hisi_ddrc.pre_cmd", "BriefDescription": "DDRC precharge commands", "PublicDescription": "DDRC precharge commands", "Unit": "hisi_sccl,ddrc",
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 20/57 perf vendor events arm64: Add some missing events for Hisi hip08 DDRC PMU
In reply to Arnaldo Carvalho de MeloFrom: John Garry <[email protected]>
Add some more missing events.
--- .../arch/arm64/hisilicon/hip08/uncore-ddrc.json | 14 ++++++++++++++ 1 file changed, 14 insertions(+)
tools/perf/pmu-events/arch/arm64/hisilicon/hip08/uncore-ddrc.json +14 -0
@@ -1,4 +1,18 @@ [ + { + "EventCode": "0x00", + "EventName": "uncore_hisi_ddrc.flux_wr", + "BriefDescription": "DDRC total write operations", + "PublicDescription": "DDRC total write operations", + "Unit": "hisi_sccl,ddrc", + }, + { + "EventCode": "0x01", + "EventName": "uncore_hisi_ddrc.flux_rd", + "BriefDescription": "DDRC total read operations", + "PublicDescription": "DDRC total read operations", + "Unit": "hisi_sccl,ddrc", + }, { "EventCode": "0x02", "EventName": "uncore_hisi_ddrc.flux_wcmd",
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 21/57 perf vendor events arm64: Add some missing events for Hisi hip08 L3C PMU
In reply to Arnaldo Carvalho de MeloFrom: John Garry <[email protected]>
Add some more missing events.
--- .../arm64/hisilicon/hip08/uncore-l3c.json | 56 +++++++++++++++++++ 1 file changed, 56 insertions(+)
tools/perf/pmu-events/arch/arm64/hisilicon/hip08/uncore-l3c.json +56 -0
@@ -34,4 +34,60 @@ "PublicDescription": "l3c precharge commands", "Unit": "hisi_sccl,l3c", }, + { + "EventCode": "0x20", + "EventName": "uncore_hisi_l3c.rd_spipe", + "BriefDescription": "Count of the number of read lines that come from this cluster of CPU core in spipe", + "PublicDescription": "Count of the number of read lines that come from this cluster of CPU core in spipe", + "Unit": "hisi_sccl,l3c", + }, + { + "EventCode": "0x21", + "EventName": "uncore_hisi_l3c.wr_spipe", + "BriefDescription": "Count of the number of write lines that come from this cluster of CPU core in spipe", + "PublicDescription": "Count of the number of write lines that come from this cluster of CPU core in spipe", + "Unit": "hisi_sccl,l3c", + }, + { + "EventCode": "0x22", + "EventName": "uncore_hisi_l3c.rd_hit_spipe", + "BriefDescription": "Count of the number of read lines that hits in spipe of this L3C", + "PublicDescription": "Count of the number of read lines that hits in spipe of this L3C", + "Unit": "hisi_sccl,l3c", + }, + { + "EventCode": "0x23", + "EventName": "uncore_hisi_l3c.wr_hit_spipe", + "BriefDescription": "Count of the number of write lines that hits in spipe of this L3C", + "PublicDescription": "Count of the number of write lines that hits in spipe of this L3C", + "Unit": "hisi_sccl,l3c", + }, + { + "EventCode": "0x29", + "EventName": "uncore_hisi_l3c.back_invalid", + "BriefDescription": "Count of the number of L3C back invalid operations", + "PublicDescription": "Count of the number of L3C back invalid operations", + "Unit": "hisi_sccl,l3c", + }, + { + "EventCode": "0x40", + "EventName": "uncore_hisi_l3c.retry_cpu", + "BriefDescription": "Count of the number of retry that L3C suppresses the CPU operations", + "PublicDescription": "Count of the number of retry that L3C suppresses the CPU operations", + "Unit": "hisi_sccl,l3c", + }, + { + "EventCode": "0x41", + "EventName": "uncore_hisi_l3c.retry_ring", + "BriefDescription": "Count of the number of retry that L3C suppresses the ring operations", + "PublicDescription": "Count of the number of retry that L3C suppresses the ring operations", + "Unit": "hisi_sccl,l3c", + }, + { + "EventCode": "0x42", + "EventName": "uncore_hisi_l3c.prefetch_drop", + "BriefDescription": "Count of the number of prefetch drops from this L3C", + "PublicDescription": "Count of the number of prefetch drops from this L3C", + "Unit": "hisi_sccl,l3c", + }, ]
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 22/57 perf vendor events arm64: Add some missing events for Hisi hip08 HHA PMU
In reply to Arnaldo Carvalho de MeloFrom: John Garry <[email protected]>
Add some more missing events.
A trivial typo is also fixed.
--- .../arm64/hisilicon/hip08/uncore-hha.json | 23 ++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-)
tools/perf/pmu-events/arch/arm64/hisilicon/hip08/uncore-hha.json +22 -1
@@ -20,6 +20,13 @@ "PublicDescription": "The number of all operations received by the HHA from another SCCL in this socket", "Unit": "hisi_sccl,hha", }, + { + "EventCode": "0x03", + "EventName": "uncore_hisi_hha.rx_ccix", + "BriefDescription": "Count of the number of operations that HHA has received from CCIX", + "PublicDescription": "Count of the number of operations that HHA has received from CCIX", + "Unit": "hisi_sccl,hha", + }, { "EventCode": "0x1c", "EventName": "uncore_hisi_hha.rd_ddr_64b",
@@ -29,7 +36,7 @@ }, { "EventCode": "0x1d", - "EventName": "uncore_hisi_hha.wr_dr_64b", + "EventName": "uncore_hisi_hha.wr_ddr_64b", "BriefDescription": "The number of write operations sent by HHA to DDRC which size is 64 bytes", "PublicDescription": "The number of write operations sent by HHA to DDRC which size is 64 bytes", "Unit": "hisi_sccl,hha",
@@ -48,4 +55,18 @@ "PublicDescription": "The number of write operations sent by HHA to DDRC which size is 128 bytes", "Unit": "hisi_sccl,hha", }, + { + "EventCode": "0x20", + "EventName": "uncore_hisi_hha.spill_num", + "BriefDescription": "Count of the number of spill operations that the HHA has sent", + "PublicDescription": "Count of the number of spill operations that the HHA has sent", + "Unit": "hisi_sccl,hha", + }, + { + "EventCode": "0x21", + "EventName": "uncore_hisi_hha.spill_success", + "BriefDescription": "Count of the number of successful spill operations that the HHA has sent", + "PublicDescription": "Count of the number of successful spill operations that the HHA has sent", + "Unit": "hisi_sccl,hha", + }, ]
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 23/57 tools arch x86: Grab a copy of the file containing the IRQ vector defines
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
We'll use it to generate a table and then convert the irq_vectors:* tracepoint 'vector' arg in things like perf trace, script, etc.
--- tools/arch/x86/include/asm/irq_vectors.h | 146 +++++++++++++++++++++++ tools/perf/check-headers.sh | 1 + 2 files changed, 147 insertions(+) create mode 100644 tools/arch/x86/include/asm/irq_vectors.h
tools/arch/x86/include/asm/irq_vectors.h new file +146 -0
@@ -0,0 +1,146 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _ASM_X86_IRQ_VECTORS_H +#define _ASM_X86_IRQ_VECTORS_H + +#include <linux/threads.h> +/* +* Linux IRQ vector layout. +* +* There are 256 IDT entries (per CPU - each entry is 8 bytes) which can +* be defined by Linux. They are used as a jump table by the CPU when a +* given vector is triggered - by a CPU-external, CPU-internal or +* software-triggered event. +* +* Linux sets the kernel code address each entry jumps to early during +* bootup, and never changes them. This is the general layout of the +* IDT entries: +* +* Vectors 0 ... 31 : system traps and exceptions - hardcoded events +* Vectors 32 ... 127 : device interrupts +* Vector 128 : legacy int80 syscall interface +* Vectors 129 ... LOCAL_TIMER_VECTOR-1 +* Vectors LOCAL_TIMER_VECTOR ... 255 : special interrupts +* +* 64-bit x86 has per CPU IDT tables, 32-bit has one shared IDT table. +* +* This file enumerates the exact layout of them: +*/ + +#define NMI_VECTOR 0x02 +#define MCE_VECTOR 0x12 + +/* +* IDT vectors usable for external interrupt sources start at 0x20. +* (0x80 is the syscall vector, 0x30-0x3f are for ISA) +*/ +#define FIRST_EXTERNAL_VECTOR 0x20 + +/* +* Reserve the lowest usable vector (and hence lowest priority) 0x20 for +* triggering cleanup after irq migration. 0x21-0x2f will still be used +* for device interrupts. +*/ +#define IRQ_MOVE_CLEANUP_VECTOR FIRST_EXTERNAL_VECTOR + +#define IA32_SYSCALL_VECTOR 0x80 + +/* +* Vectors 0x30-0x3f are used for ISA interrupts. +* round up to the next 16-vector boundary +*/ +#define ISA_IRQ_VECTOR (irq) (((FIRST_EXTERNAL_VECTOR + 16) & ~15) + irq) + +/* +* Special IRQ vectors used by the SMP architecture, 0xf0-0xff +* +* some of the following vectors are 'rare', they are merged +* into a single vector (CALL_FUNCTION_VECTOR) to save vector space. +* TLB, reschedule and local APIC vectors are performance-critical. +*/ + +#define SPURIOUS_APIC_VECTOR 0xff +/* +* Sanity check +*/ +#if ((SPURIOUS_APIC_VECTOR &0x0F ) !=0x0F ) +# error SPURIOUS_APIC_VECTOR definition error +#endif + +#define ERROR_APIC_VECTOR 0xfe +#define RESCHEDULE_VECTOR 0xfd +#define CALL_FUNCTION_VECTOR 0xfc +#define CALL_FUNCTION_SINGLE_VECTOR 0xfb +#define THERMAL_APIC_VECTOR 0xfa +#define THRESHOLD_APIC_VECTOR 0xf9 +#define REBOOT_VECTOR 0xf8 + +/* +* Generic system vector for platform specific use +*/ +#define X86_PLATFORM_IPI_VECTOR 0xf7 + +/* +* IRQ work vector: +*/ +#define IRQ_WORK_VECTOR 0xf6 + +#define UV_BAU_MESSAGE 0xf5 +#define DEFERRED_ERROR_VECTOR 0xf4 + +/* Vector on which hypervisor callbacks will be delivered */ +#define HYPERVISOR_CALLBACK_VECTOR 0xf3 + +/* Vector for KVM to deliver posted interrupt IPI */ +#ifdef CONFIG_HAVE_KVM +#define POSTED_INTR_VECTOR 0xf2 +#define POSTED_INTR_WAKEUP_VECTOR 0xf1 +#define POSTED_INTR_NESTED_VECTOR 0xf0 +#endif + +#define MANAGED_IRQ_SHUTDOWN_VECTOR 0xef + +#if IS_ENABLED (CONFIG_HYPERV ) +#define HYPERV_REENLIGHTENMENT_VECTOR 0xee +#define HYPERV_STIMER0_VECTOR 0xed +#endif + +#define LOCAL_TIMER_VECTOR 0xec + +#define NR_VECTORS 256 + +#ifdef CONFIG_X86_LOCAL_APIC +#define FIRST_SYSTEM_VECTOR LOCAL_TIMER_VECTOR +#else +#define FIRST_SYSTEM_VECTOR NR_VECTORS +#endif + +/* +* Size the maximum number of interrupts. +* +* If the irq_desc[] array has a sparse layout, we can size things +* generously - it scales up linearly with the maximum number of CPUs, +* and the maximum number of IO-APICs, whichever is higher. +* +* In other cases we size more conservatively, to not create too large +* static arrays. +*/ + +#define NR_IRQS_LEGACY 16 + +#define CPU_VECTOR_LIMIT (64 * NR_CPUS) +#define IO_APIC_VECTOR_LIMIT (32 * MAX_IO_APICS) + +#if defined(CONFIG_X86_IO_APIC ) && defined(CONFIG_PCI_MSI ) +#define NR_IRQS \ + (CPU_VECTOR_LIMIT > IO_APIC_VECTOR_LIMIT ? \ + (NR_VECTORS + CPU_VECTOR_LIMIT) : \ + (NR_VECTORS + IO_APIC_VECTOR_LIMIT)) +#elif defined(CONFIG_X86_IO_APIC ) +#define NR_IRQS (NR_VECTORS + IO_APIC_VECTOR_LIMIT) +#elif defined(CONFIG_PCI_MSI ) +#define NR_IRQS (NR_VECTORS + CPU_VECTOR_LIMIT) +#else +#define NR_IRQS NR_IRQS_LEGACY +#endif + +#endif /* _ASM_X86_IRQ_VECTORS_H */
tools/perf/check-headers.sh +1 -0
@@ -28,6 +28,7 @@ arch/x86/include/asm/disabled-features.h arch/x86/include/asm/required-features.h arch/x86/include/asm/cpufeatures.h arch/x86/include/asm/inat_types.h +arch/x86/include/asm/irq_vectors.h arch/x86/include/asm/msr-index.h arch/x86/include/uapi/asm/prctl.h arch/x86/lib/x86-opcode-map.txt
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 24/57 libbeauty: Add a generator for x86's IRQ vectors -> strings
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
We'll wire this up with the 'vector' arg in irq_vectors:*, etc:
Just run it straight away and check what it produces:
$ tools/perf/trace/beauty/tracepoints/x86_irq_vectors.sh
static const char *x86_irq_vectors[] = {
[0x02] = "NMI",
[0x12] = "MCE",
[0x20] = "IRQ_MOVE_CLEANUP",
[0x80] = "IA32_SYSCALL",
[0xec] = "LOCAL_TIMER",
[0xed] = "HYPERV_STIMER0",
[0xee] = "HYPERV_REENLIGHTENMENT",
[0xef] = "MANAGED_IRQ_SHUTDOWN",
[0xf0] = "POSTED_INTR_NESTED",
[0xf1] = "POSTED_INTR_WAKEUP",
[0xf2] = "POSTED_INTR",
[0xf3] = "HYPERVISOR_CALLBACK",
[0xf4] = "DEFERRED_ERROR",
[0xf6] = "IRQ_WORK",
[0xf7] = "X86_PLATFORM_IPI",
[0xf8] = "REBOOT",
[0xf9] = "THRESHOLD_APIC",
[0xfa] = "THERMAL_APIC",
[0xfb] = "CALL_FUNCTION_SINGLE",
[0xfc] = "CALL_FUNCTION",
[0xfd] = "RESCHEDULE",
[0xfe] = "ERROR_APIC",
[0xff] = "SPURIOUS_APIC",
};
$--- .../beauty/tracepoints/x86_irq_vectors.sh | 27 +++++++++++++++++++ 1 file changed, 27 insertions(+) create mode 100755 tools/perf/trace/beauty/tracepoints/x86_irq_vectors.sh
tools/perf/trace/beauty/tracepoints/x86_irq_vectors.sh new file +27 -0
@@ -0,0 +1,27 @@ +#!/bin/sh +# SPDX-License-Identifier: LGPL-2.1 +# (C) 2019, Arnaldo Carvalho de Melo <[email protected]> + +if [ $# -ne 1 ] ; then + arch_x86_header_dir=tools/arch/x86/include/asm/ +else + arch_x86_header_dir=$1 +fi + +x86_irq_vectors=${arch_x86_header_dir}/irq_vectors.h + +# FIRST_EXTERNAL_VECTOR is not that useful, find what is its number +# and then replace whatever is using it and that is useful, which at +# the time of writing of this script was: IRQ_MOVE_CLEANUP_VECTOR. + +first_external_regex='^#define[[:space:]]+FIRST_EXTERNAL_VECTOR[[:space:]]+(0x[[:xdigit:]]+)$' +first_external_vector=$(egrep ${first_external_regex} ${x86_irq_vectors} | sed -r "s/${first_external_regex}/\1/g") + +printf "static const char *x86_irq_vectors[] = {\n" +regex='^#define[[:space:]]+([[:alnum:]_]+)_VECTOR[[:space:]]+(0x[[:xdigit:]]+)$' +sed -r "s/FIRST_EXTERNAL_VECTOR/${first_external_vector}/g" ${x86_irq_vectors} | \ +egrep ${regex} | \ + sed -r "s/${regex}/\2 \1/g" | sort -n | \ + xargs printf "\t[%s] = \"%s\",\n" +printf "};\n\n" +
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 25/57 libbeauty: Hook up the x86 irq_vectors table generator
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
I.e. after running:
$ make -C tools/perf O=/tmp/build/perf
We end up with:
$ cat /tmp/build/perf/trace/beauty/generated/x86_arch_irq_vectors_array.c
static const char *x86_irq_vectors[] = {
[0x02] = "NMI",
[0x12] = "MCE",
[0x20] = "IRQ_MOVE_CLEANUP",
[0x80] = "IA32_SYSCALL",
[0xec] = "LOCAL_TIMER",
[0xed] = "HYPERV_STIMER0",
[0xee] = "HYPERV_REENLIGHTENMENT",
[0xef] = "MANAGED_IRQ_SHUTDOWN",
[0xf0] = "POSTED_INTR_NESTED",
[0xf1] = "POSTED_INTR_WAKEUP",
[0xf2] = "POSTED_INTR",
[0xf3] = "HYPERVISOR_CALLBACK",
[0xf4] = "DEFERRED_ERROR",
[0xf6] = "IRQ_WORK",
[0xf7] = "X86_PLATFORM_IPI",
[0xf8] = "REBOOT",
[0xf9] = "THRESHOLD_APIC",
[0xfa] = "THERMAL_APIC",
[0xfb] = "CALL_FUNCTION_SINGLE",
[0xfc] = "CALL_FUNCTION",
[0xfd] = "RESCHEDULE",
[0xfe] = "ERROR_APIC",
[0xff] = "SPURIOUS_APIC",
};
$Now its just a matter of using it, associating it to tracepoint arguments named 'vector', all of which can be correctly used with this table, for int args.
At some point we should move tools/perf/trace/beauty to tools/beauty/, so that it can be used more generally and even made available externally like libbpf, libperf, libtraceevent, etc.
--- tools/perf/Makefile.perf | 8 ++++++++ 1 file changed, 8 insertions(+)
tools/perf/Makefile.perf +8 -0
@@ -546,6 +546,12 @@ x86_arch_prctl_code_tbl := $(srctree)/tools/perf/trace/beauty/x86_arch_prctl.sh $(x86_arch_prctl_code_array): $(x86_arch_asm_uapi_dir)/prctl.h $(x86_arch_prctl_code_tbl) $(Q)$(SHELL) '$(x86_arch_prctl_code_tbl)' $(x86_arch_asm_uapi_dir) > $@ +x86_arch_irq_vectors_array := $(beauty_outdir)/x86_arch_irq_vectors_array.c +x86_arch_irq_vectors_tbl := $(srctree)/tools/perf/trace/beauty/tracepoints/x86_irq_vectors.sh + +$(x86_arch_irq_vectors_array): $(x86_arch_asm_dir)/irq_vectors.h $(x86_arch_irq_vectors_tbl) + $(Q)$(SHELL) '$(x86_arch_irq_vectors_tbl)' $(x86_arch_asm_dir) > $@ + x86_arch_MSRs_array := $(beauty_outdir)/x86_arch_MSRs_array.c x86_arch_MSRs_tbl := $(srctree)/tools/perf/trace/beauty/tracepoints/x86_msr.sh
@@ -686,6 +692,7 @@ prepare: $(OUTPUT)PERF-VERSION-FILE $(OUTPUT)common-cmds.h archheaders $(drm_ioc $(perf_ioctl_array) \ $(prctl_option_array) \ $(usbdevfs_ioctl_array) \ + $(x86_arch_irq_vectors_array) \ $(x86_arch_MSRs_array) \ $(x86_arch_prctl_code_array) \ $(rename_flags_array) \
@@ -991,6 +998,7 @@ clean:: $(LIBTRACEEVENT)-clean $(LIBAPI)-clean $(LIBBPF)-clean $(LIBSUBCMD)-clea $(OUTPUT)$(perf_ioctl_array) \ $(OUTPUT)$(prctl_option_array) \ $(OUTPUT)$(usbdevfs_ioctl_array) \ + $(OUTPUT)$(x86_arch_irq_vectors_array) \ $(OUTPUT)$(x86_arch_MSRs_array) \ $(OUTPUT)$(x86_arch_prctl_code_array) \ $(OUTPUT)$(rename_flags_array) \
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 26/57 libbeauty: Add a strarray__scnprintf_suffix() method
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
In some cases, like with x86 IRQ vectors, the common part in names is at the end, so a suffix, add a scnprintf function for that.
--- tools/perf/builtin-trace.c | 14 ++++++++++++++ tools/perf/trace/beauty/beauty.h | 1 + 2 files changed, 15 insertions(+)
tools/perf/builtin-trace.c +14 -0
@@ -423,6 +423,20 @@ static struct evsel *perf_evsel__raw_syscall_newtp(const char *direction, void * ({struct syscall_tp *fields = evsel->priv; \ fields->name.pointer(&fields->name, sample); }) +size_t strarray__scnprintf_suffix (struct strarray *sa,char *bf,size_t size,const char *intfmt,bool show_suffix,int val) +{ +int idx = val - sa->offset; + +if (idx <0 || idx >= sa->nr_entries || sa->entries[idx] ==NULL ) { +size_t printed =scnprintf (bf, size, intfmt, val); +if (show_suffix) + printed +=scnprintf (bf + printed, size - printed," /* %s??? */" , sa->prefix); +return printed; + } + +return scnprintf (bf, size,"%s%s" , sa->entries[idx], show_suffix ? sa->prefix :"" ); +} +size_t strarray__scnprintf (struct strarray *sa,char *bf,size_t size,const char *intfmt,bool show_prefix,int val) {int idx = val - sa->offset;
tools/perf/trace/beauty/beauty.h +1 -0
@@ -28,6 +28,7 @@ struct strarray { }size_t strarray__scnprintf (struct strarray *sa,char *bf,size_t size,const char *intfmt,bool show_prefix,int val); +size_t strarray__scnprintf_suffix (struct strarray *sa,char *bf,size_t size,const char *intfmt,bool show_suffix,int val);size_t strarray__scnprintf_flags (struct strarray *sa,char *bf,size_t size,bool show_prefix,unsigned long flags);bool strarray__strtoul (struct strarray *sa,char *bf,size_t size,u64 *ret);
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 27/57 perf trace beauty: Add the glue for the autogenerated x86 IRQ vector array
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
We need to wrap this autogenerated string array with the strarray__scnprintf() formatter and the strarray__strotul() lookup method, do it.
--- tools/perf/trace/beauty/beauty.h | 6 ++++ tools/perf/trace/beauty/tracepoints/Build | 1 + .../beauty/tracepoints/x86_irq_vectors.c | 29 +++++++++++++++++++ 3 files changed, 36 insertions(+) create mode 100644 tools/perf/trace/beauty/tracepoints/x86_irq_vectors.c
tools/perf/trace/beauty/beauty.h +6 -0
@@ -122,6 +122,12 @@ unsigned long syscall_arg__val(struct syscall_arg *arg, u8 idx);size_t syscall_arg__scnprintf_strarray_flags (char *bf,size_t size,struct syscall_arg *arg);#define SCA_STRARRAY_FLAGS syscall_arg__scnprintf_strarray_flags +size_t syscall_arg__scnprintf_x86_irq_vectors (char *bf,size_t size,struct syscall_arg *arg); +#define SCA_X86_IRQ_VECTORS syscall_arg__scnprintf_x86_irq_vectors + +bool syscall_arg__strtoul_x86_irq_vectors (char *bf,size_t size,struct syscall_arg *arg,u64 *ret); +#define STUL_X86_IRQ_VECTORS syscall_arg__strtoul_x86_irq_vectors +size_t syscall_arg__scnprintf_x86_MSR (char *bf,size_t size,struct syscall_arg *arg);#define SCA_X86_MSR syscall_arg__scnprintf_x86_MSR
tools/perf/trace/beauty/tracepoints/Build +1 -0
@@ -1 +1,2 @@ +perf-y += x86_irq_vectors.o perf-y += x86_msr.o
tools/perf/trace/beauty/tracepoints/x86_irq_vectors.c new file +29 -0
@@ -0,0 +1,29 @@ +// SPDX-License-Identifier: LGPL-2.1 +/* +* trace/beauty/x86_irq_vectors.c +* +* Copyright (C) 2019, Red Hat Inc, Arnaldo Carvalho de Melo <[email protected]> +*/ + +#include "trace/beauty/beauty.h" + +#include "trace/beauty/generated/x86_arch_irq_vectors_array.c" + +static DEFINE_STRARRAY(x86_irq_vectors, "_VECTOR"); + +static size_t x86_irq_vectors__scnprintf (unsigned long vector,char *bf,size_t size,bool show_prefix) +{ +return strarray__scnprintf_suffix (&strarray__x86_irq_vectors, bf, size,"%#x" , show_prefix, vector); +} + +size_t syscall_arg__scnprintf_x86_irq_vectors (char *bf,size_t size,struct syscall_arg *arg) +{ +unsigned long vector = arg->val; + +return x86_irq_vectors__scnprintf (vector, bf, size, arg->show_string_prefix); +} + +bool syscall_arg__strtoul_x86_irq_vectors (char *bf,size_t size,struct syscall_arg *arg__maybe_unused ,u64 *ret) +{ +return strarray__strtoul (&strarray__x86_irq_vectors, bf, size, ret); +}
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 28/57 perf trace: Hook the 'vec' tracepoint argument with the x86 IRQ vectors scnprintf/strtoul
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
Ended up only being useful when filtering multiple irq_vectors tracepoints, as we end up having a tracepoint for each of the entries, i.e.:
This will always come with the "RESCHEDULE_VECTOR" in the 'vector' arg:
# perf trace --max-events 8 -e irq_vectors:reschedule*
0.000 cc1/29067 irq_vectors:reschedule_entry(vector: RESCHEDULE)
0.004 cc1/29067 irq_vectors:reschedule_exit(vector: RESCHEDULE)
0.553 cc1/29067 irq_vectors:reschedule_entry(vector: RESCHEDULE)
0.556 cc1/29067 irq_vectors:reschedule_exit(vector: RESCHEDULE)
1.182 cc1/29067 irq_vectors:reschedule_entry(vector: RESCHEDULE)
1.185 cc1/29067 irq_vectors:reschedule_exit(vector: RESCHEDULE)
1.203 :29052/29052 irq_vectors:reschedule_entry(vector: RESCHEDULE)
1.206 :29052/29052 irq_vectors:reschedule_exit(vector: RESCHEDULE)
#While filtering that value will produce nothing:
# perf trace --max-events 8 -e irq_vectors:reschedule* --filter="vector != RESCHEDULE" ^C#
Maybe it'll be useful for those other tracepoints:
# perf list irq_vectors:vector_*
List of pre-defined events (to be used in -e):
irq_vectors:vector_activate [Tracepoint event]
irq_vectors:vector_alloc [Tracepoint event]
irq_vectors:vector_alloc_managed [Tracepoint event]
irq_vectors:vector_clear [Tracepoint event]
irq_vectors:vector_config [Tracepoint event]
irq_vectors:vector_deactivate [Tracepoint event]
irq_vectors:vector_free_moved [Tracepoint event]
irq_vectors:vector_reserve [Tracepoint event]
irq_vectors:vector_reserve_managed [Tracepoint event]
irq_vectors:vector_setup [Tracepoint event]
irq_vectors:vector_teardown [Tracepoint event]
irq_vectors:vector_update [Tracepoint event]#
But since we have it done, keep it.
This at least served to teach me that all those irq vectors have a entry and an exit tracepoint that I can then use just like with raw_syscalls:sys_{enter,exit}, i.e. pair them, use just a trace__irq_vectors_entry() + trace__irq_vectors_exit() and use the 'vector' arg as I use the 'syscall id' one for syscalls.
Then the default for 'perf trace' will include irq_vectors in addition to syscalls.
--- tools/perf/builtin-trace.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tools/perf/builtin-trace.c +2 -1
@@ -1528,7 +1528,8 @@ static int syscall__alloc_arg_fmts(struct syscall *sc, int nr_args) }static struct syscall_arg_fmt syscall_arg_fmts__by_name[] = { - { .name ="msr" , .scnprintf =SCA_X86_MSR , .strtoul =STUL_X86_MSR , } + { .name ="msr" , .scnprintf =SCA_X86_MSR , .strtoul =STUL_X86_MSR , }, + { .name ="vector" , .scnprintf =SCA_X86_IRQ_VECTORS , .strtoul =STUL_X86_IRQ_VECTORS , }, };static int syscall_arg_fmt__cmp (const void *name,const void *fmtp)
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 29/57 perf trace: Show error message when not finding a field used in a filter expression
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
It was there, but as pr_debug(), make it pr_err() so that we can see it without -v:
# trace -e syscalls:*lseek --filter="whenc==SET" sleep 1 "whenc" not found in "syscalls:sys_enter_lseek", can't set filter "whenc==SET" #
--- tools/perf/builtin-trace.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
tools/perf/builtin-trace.c +2 -2
@@ -3611,8 +3611,8 @@ static int trace__expand_filter(struct trace *trace __maybe_unused, struct evsel fmt =perf_evsel__syscall_arg_fmt (evsel, arg);if (fmt ==NULL ) { -pr_debug ("\"%s\" not found in \"%s\", can't set filter \"%s\"\n" , - arg, evsel->name, evsel->filter); +pr_err ("\"%s\" not found in \"%s\", can't set filter \"%s\"\n" , + arg, evsel->name, evsel->filter);return -1 ; }
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 30/57 perf trace: Introduce accessors to trace specific evsel->priv
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
We're using evsel->priv in syscalls:sys_{enter,exit}_SYSCALL and in raw_syscalls:sys_{enter,exit} to cache the offset of the common fields, the multiplexor id/syscall_id in the sys_enter case and syscall_id + ret for sys_exit.
And for the rest of the tracepoints we use it to have a syscall_arg_fmt array to have scnprintf/strtoul for tracepoint args.
So we better clearly mark them with accessors so that we can move to having a 'struct evsel_trace' struct for all 'perf trace' specific evsel->priv usage.
--- tools/perf/builtin-trace.c | 43 ++++++++++++++++++++++++++++---------- 1 file changed, 32 insertions(+), 11 deletions(-)
tools/perf/builtin-trace.c +32 -11
@@ -285,6 +285,27 @@ struct syscall_tp { }; }; +/* +* Used with raw_syscalls:sys_{enter,exit} and with the +* syscalls:sys_{enter,exit}_SYSCALL tracepoints +*/ +static inline struct syscall_tp *__evsel__syscall_tp (struct evsel *evsel) +{ +struct syscall_tp *sc = evsel->priv; + +return sc; +} + +/* +* Used with all the other tracepoints. +*/ +static inline struct syscall_arg_fmt *__evsel__syscall_arg_fmt (struct evsel *evsel) +{ +struct syscall_arg_fmt *fmt = evsel->priv; + +return fmt; +} +static int perf_evsel__init_tp_uint_field (struct evsel *evsel,struct tp_field *field,const char *name)
@@ -298,7 +319,7 @@ static int perf_evsel__init_tp_uint_field(struct evsel *evsel, }#define perf_evsel__init_sc_tp_uint_field (evsel, name) \ - ({ struct syscall_tp *sc = evsel->priv;\ + ({ struct syscall_tp *sc = __evsel__syscall_tp(evsel);\ perf_evsel__init_tp_uint_field(evsel, &sc->name, #name); }) static int perf_evsel__init_tp_ptr_field(struct evsel *evsel,
@@ -314,7 +335,7 @@ static int perf_evsel__init_tp_ptr_field(struct evsel *evsel, }#define perf_evsel__init_sc_tp_ptr_field (evsel, name) \ - ({ struct syscall_tp *sc = evsel->priv;\ + ({ struct syscall_tp *sc = __evsel__syscall_tp(evsel);\ perf_evsel__init_tp_ptr_field(evsel, &sc->name, #name); })static void evsel__delete_priv (struct evsel *evsel)
@@ -364,14 +385,14 @@ static int perf_evsel__init_augmented_syscall_tp(struct evsel *evsel, struct evsstatic int perf_evsel__init_augmented_syscall_tp_args (struct evsel *evsel) { -struct syscall_tp *sc = evsel->priv; +struct syscall_tp *sc =__evsel__syscall_tp (evsel);return __tp_field__init_ptr (&sc->args, sc->id.offset +sizeof (u64)); }static int perf_evsel__init_augmented_syscall_tp_ret (struct evsel *evsel) { -struct syscall_tp *sc = evsel->priv; +struct syscall_tp *sc =__evsel__syscall_tp (evsel);return __tp_field__init_uint (&sc->ret,sizeof (u64), sc->id.offset +sizeof (u64), evsel->needs_swap); }
@@ -416,11 +437,11 @@ static struct evsel *perf_evsel__raw_syscall_newtp(const char *direction, void * }#define perf_evsel__sc_tp_uint (evsel, name, sample) \ - ({ struct syscall_tp *fields = evsel->priv; \ + ({ struct syscall_tp *fields = __evsel__syscall_tp(evsel); \ fields->name.integer(&fields->name, sample); })#define perf_evsel__sc_tp_ptr (evsel, name, sample) \ - ({ struct syscall_tp *fields = evsel->priv; \ + ({ struct syscall_tp *fields = __evsel__syscall_tp(evsel); \ fields->name.pointer(&fields->name, sample); })size_t strarray__scnprintf_suffix (struct strarray *sa,char *bf,size_t size,const char *intfmt,bool show_suffix,int val)
@@ -2518,7 +2539,7 @@ static size_t trace__fprintf_tp_fields(struct trace *trace, struct evsel *evsel,char bf[2048 ];size_t size =sizeof (bf);struct tep_format_field *field = evsel->tp_format->format.fields; -struct syscall_arg_fmt *arg = evsel->priv; +struct syscall_arg_fmt *arg =__evsel__syscall_arg_fmt (evsel);size_t printed =0 ;unsigned long val;u8 bit =1 ;
@@ -3557,7 +3578,7 @@ static int ordered_events__deliver_event(struct ordered_events *oe,static struct syscall_arg_fmt *perf_evsel__syscall_arg_fmt (struct evsel *evsel,char *arg) {struct tep_format_field *field; -struct syscall_arg_fmt *fmt = evsel->priv; +struct syscall_arg_fmt *fmt =__evsel__syscall_arg_fmt (evsel);if (evsel->tp_format ==NULL || fmt ==NULL )return NULL ;
@@ -4315,12 +4336,12 @@ static int evlist__set_syscall_tp_fields(struct evlist *evlist)return -1 ;if (!strncmp (evsel->tp_format->name,"sys_enter_" ,10 )) { -struct syscall_tp *sc = evsel->priv; +struct syscall_tp *sc =__evsel__syscall_tp (evsel);if (__tp_field__init_ptr (&sc->args, sc->id.offset +sizeof (u64)))return -1 ; }else if (!strncmp (evsel->tp_format->name,"sys_exit_" ,9 )) { -struct syscall_tp *sc = evsel->priv; +struct syscall_tp *sc =__evsel__syscall_tp (evsel);if (__tp_field__init_uint (&sc->ret,sizeof (u64), sc->id.offset +sizeof (u64), evsel->needs_swap))return -1 ;
@@ -4856,7 +4877,7 @@ int cmd_trace(int argc, const char **argv) init_augmented_syscall_tp:if (perf_evsel__init_augmented_syscall_tp (evsel, evsel)) goto out; - sc = evsel->priv; + sc =__evsel__syscall_tp (evsel);/* * For now with BPF raw_augmented we hook into * raw_syscalls:sys_enter and there we get all
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 31/57 perf trace: Hide evsel->access further, simplify code
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
Next step will be to have a 'struct evsel_trace' to allow for handling the syscalls tracepoints via the strace-like code while reusing parts of that code with the other tracepoints, where we don't have things like the 'syscall_nr' or 'ret' ((raw_)?syscalls:sys_{enter,exit}(_SYSCALL)?) args that we want to cache offsets and have been using evsel->priv for that, while for the other tracepoints we'll have just an array of 'struct syscall_arg_fmt' (i.e. ->scnprint() for number->string and ->strtoul() string->number conversions and other state those functions need).
--- tools/perf/builtin-trace.c | 57 +++++++++++++++++++++----------------- 1 file changed, 31 insertions(+), 26 deletions(-)
tools/perf/builtin-trace.c +31 -26
@@ -296,6 +296,15 @@ static inline struct syscall_tp *__evsel__syscall_tp(struct evsel *evsel)return sc; } +static struct syscall_tp *evsel__syscall_tp (struct evsel *evsel) +{ +if (evsel->priv ==NULL ) { + evsel->priv =zalloc (sizeof (struct syscall_tp )); + } + +return __evsel__syscall_tp (evsel); +} +/* * Used with all the other tracepoints. */
@@ -306,6 +315,15 @@ static inline struct syscall_arg_fmt *__evsel__syscall_arg_fmt(struct evsel *evsreturn fmt; } +static struct syscall_arg_fmt *evsel__syscall_arg_fmt (struct evsel *evsel) +{ +if (evsel->priv ==NULL ) { + evsel->priv =calloc (evsel->tp_format->format.nr_fields,sizeof (struct syscall_arg_fmt )); + } + +return __evsel__syscall_arg_fmt (evsel); +} +static int perf_evsel__init_tp_uint_field (struct evsel *evsel,struct tp_field *field,const char *name)
@@ -346,41 +364,34 @@ static void evsel__delete_priv(struct evsel *evsel)static int perf_evsel__init_syscall_tp (struct evsel *evsel) { -struct syscall_tp *sc = evsel->priv =malloc (sizeof (struct syscall_tp )); +struct syscall_tp *sc =evsel__syscall_tp (evsel); -if (evsel->priv !=NULL ) { +if (sc !=NULL ) {if (perf_evsel__init_tp_uint_field (evsel, &sc->id,"__syscall_nr" ) &&perf_evsel__init_tp_uint_field (evsel, &sc->id,"nr" )) - goto out_delete; +return -ENOENT ;return 0 ; }return -ENOMEM ; -out_delete: -zfree (&evsel->priv); -return -ENOENT ; }static int perf_evsel__init_augmented_syscall_tp (struct evsel *evsel,struct evsel *tp) { -struct syscall_tp *sc = evsel->priv =malloc (sizeof (struct syscall_tp )); +struct syscall_tp *sc =evsel__syscall_tp (evsel); -if (evsel->priv !=NULL ) { +if (sc !=NULL ) {struct tep_format_field *syscall_id =perf_evsel__field (tp,"id" );if (syscall_id ==NULL ) syscall_id =perf_evsel__field (tp,"__syscall_nr" ); -if (syscall_id ==NULL ) - goto out_delete; -if (__tp_field__init_uint (&sc->id, syscall_id->size, syscall_id->offset, evsel->needs_swap)) - goto out_delete; +if (syscall_id ==NULL || +__tp_field__init_uint (&sc->id, syscall_id->size, syscall_id->offset, evsel->needs_swap)) +return -EINVAL ;return 0 ; }return -ENOMEM ; -out_delete: -zfree (&evsel->priv); -return -EINVAL ; }static int perf_evsel__init_augmented_syscall_tp_args (struct evsel *evsel)
@@ -399,20 +410,15 @@ static int perf_evsel__init_augmented_syscall_tp_ret(struct evsel *evsel)static int perf_evsel__init_raw_syscall_tp (struct evsel *evsel,void *handler) { - evsel->priv =malloc (sizeof (struct syscall_tp )); -if (evsel->priv !=NULL ) { +if (evsel__syscall_tp (evsel) !=NULL ) {if (perf_evsel__init_sc_tp_uint_field (evsel, id)) - goto out_delete; +return -ENOENT ; evsel->handler = handler;return 0 ; }return -ENOMEM ; - -out_delete: -zfree (&evsel->priv); -return -ENOENT ; }static struct evsel *perf_evsel__raw_syscall_newtp (const char *direction,void *handler)
@@ -1690,11 +1696,10 @@ static int trace__read_syscall_info(struct trace *trace, int id)static int perf_evsel__init_tp_arg_scnprintf (struct evsel *evsel) { -int nr_args = evsel->tp_format->format.nr_fields; +struct syscall_arg_fmt *fmt =evsel__syscall_arg_fmt (evsel); - evsel->priv =calloc (nr_args,sizeof (struct syscall_arg_fmt )); -if (evsel->priv !=NULL ) { -syscall_arg_fmt__init_array (evsel->priv, evsel->tp_format->format.fields); +if (fmt !=NULL ) { +syscall_arg_fmt__init_array (fmt, evsel->tp_format->format.fields);return 0 ; }
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 32/57 perf trace: Introduce 'struct evsel__trace' for evsel->priv needs
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
For syscalls we need to cache the 'syscall_id' and 'ret' field offsets but as well have a pointer to the syscall_fmt_arg array for the fields, so that we can expand strings in filter expressions, so introduce a 'struct evsel_trace' to have in evsel->priv that allows for that.
--- tools/perf/builtin-trace.c | 54 +++++++++++++++++++++++++++++++++----- 1 file changed, 48 insertions(+), 6 deletions(-)
tools/perf/builtin-trace.c +48 -6
@@ -285,21 +285,47 @@ struct syscall_tp { }; }; +/* +* The evsel->priv as used by 'perf trace' +* sc: for raw_syscalls:sys_{enter,exit} and syscalls:sys_{enter,exit}_SYSCALLNAME +* fmt: for all the other tracepoints +*/ +struct evsel_trace { +struct syscall_tp sc; +struct syscall_arg_fmt *fmt; +}; + +static struct evsel_trace *evsel_trace__new (void ) +{ +return zalloc (sizeof (struct evsel_trace )); +} + +static void evsel_trace__delete (struct evsel_trace *et) +{ +if (et ==NULL ) +return ; + +zfree (&et->fmt); +free (et); +} +/* * Used with raw_syscalls:sys_{enter,exit} and with the * syscalls:sys_{enter,exit}_SYSCALL tracepoints */ static inline struct syscall_tp *__evsel__syscall_tp (struct evsel *evsel) { -struct syscall_tp *sc = evsel->priv; +struct evsel_trace *et = evsel->priv; -return sc; +return &et->sc; }static struct syscall_tp *evsel__syscall_tp (struct evsel *evsel) {if (evsel->priv ==NULL ) { - evsel->priv =zalloc (sizeof (struct syscall_tp )); + evsel->priv =evsel_trace__new (); +if (evsel->priv ==NULL ) +return NULL ; }return __evsel__syscall_tp (evsel);
@@ -310,18 +336,34 @@ static struct syscall_tp *evsel__syscall_tp(struct evsel *evsel)*/ static inline struct syscall_arg_fmt *__evsel__syscall_arg_fmt (struct evsel *evsel) { -struct syscall_arg_fmt *fmt = evsel->priv; +struct evsel_trace *et = evsel->priv; -return fmt; +return et->fmt; }static struct syscall_arg_fmt *evsel__syscall_arg_fmt (struct evsel *evsel) { +struct evsel_trace *et = evsel->priv; +if (evsel->priv ==NULL ) { - evsel->priv =calloc (evsel->tp_format->format.nr_fields,sizeof (struct syscall_arg_fmt )); + et = evsel->priv =evsel_trace__new (); + +if (et ==NULL ) +return NULL ; + } + +if (et->fmt ==NULL ) { + et->fmt =calloc (evsel->tp_format->format.nr_fields,sizeof (struct syscall_arg_fmt )); +if (et->fmt ==NULL ) + goto out_delete; }return __evsel__syscall_arg_fmt (evsel); + +out_delete: +evsel_trace__delete (evsel->priv); + evsel->priv =NULL ; +return NULL ; } static int perf_evsel__init_tp_uint_field(struct evsel *evsel,
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 33/57 perf trace: Initialize evsel_trace->fmt for syscalls:sys_enter_* tracepoints
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
From the syscall_fmts->arg entries for formatting strace-like syscalls.
This is when resolving the string "whence" on a filter expression for the syscalls:sys_enter_lseek:
Breakpoint 3, perf_evsel__syscall_arg_fmt (evsel=0xc91ed0, arg=0x7fffffff7cd0 "whence") at builtin-trace.c:3626
3626 {
(gdb) n
3628 struct syscall_arg_fmt *fmt = __evsel__syscall_arg_fmt(evsel);
(gdb) n
3630 if (evsel->tp_format == NULL || fmt == NULL)
(gdb) n
3633 for (field = evsel->tp_format->format.fields; field; field = field->next, ++fmt)
(gdb) n
3634 if (strcmp(field->name, arg) == 0)
(gdb) p field->name
$3 = 0xc945e0 "__syscall_nr"
(gdb) n
3633 for (field = evsel->tp_format->format.fields; field; field = field->next, ++fmt)
(gdb) p *fmt
$4 = {scnprintf = 0x0, strtoul = 0x0, mask_val = 0x0, parm = 0x0, name = 0x0, nr_entries = 0, show_zero = false}
(gdb) n
3634 if (strcmp(field->name, arg) == 0)
(gdb) p field->name
$5 = 0xc94690 "fd"
(gdb) n
3633 for (field = evsel->tp_format->format.fields; field; field = field->next, ++fmt)
(gdb) n
3634 if (strcmp(field->name, arg) == 0)
(gdb) n
3633 for (field = evsel->tp_format->format.fields; field; field = field->next, ++fmt)
(gdb) n
3634 if (strcmp(field->name, arg) == 0)
(gdb) p *fmt
$9 = {scnprintf = 0x489be2 <syscall_arg__scnprintf_strarray>, strtoul = 0x0, mask_val = 0x0, parm = 0xa2da80 <strarray.whences>, name = 0x0,
nr_entries = 0, show_zero = false}
(gdb) p field->name
$10 = 0xc947b0 "whence"
(gdb) p fmt->parm
$11 = (void *) 0xa2da80 <strarray.whences>
(gdb) p *(struct strarray *)fmt->parm
$12 = {offset = 0, nr_entries = 5, prefix = 0x724d37 "SEEK_", entries = 0xa2da40 <whences>}
(gdb) p (struct strarray *)fmt->parm)->entries
Junk after end of expression.
(gdb) p ((struct strarray *)fmt->parm)->entries
$13 = (const char **) 0xa2da40 <whences>
(gdb) p ((struct strarray *)fmt->parm)->entries[0]
$14 = 0x724d21 "SET"
(gdb) p ((struct strarray *)fmt->parm)->entries[1]
$15 = 0x724d25 "CUR"
(gdb) p ((struct strarray *)fmt->parm)->entries[2]
$16 = 0x724d29 "END"
(gdb) p ((struct strarray *)fmt->parm)->entries[2]
$17 = 0x724d29 "END"
(gdb) p ((struct strarray *)fmt->parm)->entries[3]
$18 = 0x724d2d "DATA"
(gdb) p ((struct strarray *)fmt->parm)->entries[4]
$19 = 0x724d32 "HOLE"
(gdb)--- tools/perf/builtin-trace.c | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+)
tools/perf/builtin-trace.c +23 -0
@@ -4366,6 +4366,25 @@ static void evlist__set_default_evsel_handler(struct evlist *evlist, void *handl } } +static void evsel__set_syscall_arg_fmt (struct evsel *evsel,const char *name) +{ +struct syscall_arg_fmt *fmt =evsel__syscall_arg_fmt (evsel); + +if (fmt) { +struct syscall_fmt *scfmt =syscall_fmt__find (name); + +if (scfmt) { +int skip =0 ; + +if (strcmp (evsel->tp_format->format.fields->name,"__syscall_nr" ) ==0 || +strcmp (evsel->tp_format->format.fields->name,"nr" ) ==0 ) + ++skip; + +memcpy (fmt + skip, scfmt->arg, (evsel->tp_format->format.nr_fields - skip) *sizeof (*fmt)); + } + } +} +static int evlist__set_syscall_tp_fields (struct evlist *evlist) {struct evsel *evsel;
@@ -4387,11 +4406,15 @@ static int evlist__set_syscall_tp_fields(struct evlist *evlist)if (__tp_field__init_ptr (&sc->args, sc->id.offset +sizeof (u64)))return -1 ; + +evsel__set_syscall_arg_fmt (evsel, evsel->tp_format->name +sizeof ("sys_enter_" ) -1 ); }else if (!strncmp (evsel->tp_format->name,"sys_exit_" ,9 )) {struct syscall_tp *sc =__evsel__syscall_tp (evsel);if (__tp_field__init_uint (&sc->ret,sizeof (u64), sc->id.offset +sizeof (u64), evsel->needs_swap))return -1 ; + +evsel__set_syscall_arg_fmt (evsel, evsel->tp_format->name +sizeof ("sys_exit_" ) -1 ); } }
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 34/57 perf scripting engines: Iterate on tep event arrays directly
In reply to Arnaldo Carvalho de MeloFrom: "Steven Rostedt (VMware)" <[email protected]>
Instead of calling a useless (and broken) helper function to get the next event of a tep event array, just get the array directly and iterate over it.
Note, the broken part was from trace_find_next_event() which after this will no longer be used, and can be removed.
Committer notes:
This fixes a segfault when generating python scripts from perf.data files with multiple tracepoint events, i.e. the following use case is fixed by this patch:
# perf record -e sched:* sleep 1 [ perf record: Woken up 31 times to write data ] [ perf record: Captured and wrote 0.031 MB perf.data (9 samples) ] # perf script -g python Segmentation fault (core dumped) #
--- tools/perf/util/scripting-engines/trace-event-perl.c | 8 ++++++-- tools/perf/util/scripting-engines/trace-event-python.c | 9 +++++++-- 2 files changed, 13 insertions(+), 4 deletions(-)
tools/perf/util/scripting-engines/trace-event-perl.c +6 -2
@@ -539,10 +539,11 @@ static int perl_stop_script(void)static int perl_generate_script (struct tep_handle *pevent,const char *outfile) { +int i, not_first, count, nr_events; +struct tep_event **all_events;struct tep_event *event =NULL ;struct tep_format_field *f;char fname[PATH_MAX ]; -int not_first, count;FILE *ofp;sprintf (fname,"%s.pl" , outfile);
@@ -603,8 +604,11 @@ sub print_backtrace\n\ }\n\n\ "); + nr_events =tep_get_events_count (pevent); + all_events =tep_list_events (pevent,TEP_EVENT_SORT_ID ); -while ((event =trace_find_next_event (pevent, event))) { +for (i =0 ; all_events && i < nr_events; i++) { + event = all_events[i];fprintf (ofp,"sub %s::%s\n{\n" , event->system, event->name);fprintf (ofp,"\tmy (" );
tools/perf/util/scripting-engines/trace-event-python.c +7 -2
@@ -1687,10 +1687,11 @@ static int python_stop_script(void)static int python_generate_script (struct tep_handle *pevent,const char *outfile) { +int i, not_first, count, nr_events; +struct tep_event **all_events;struct tep_event *event =NULL ;struct tep_format_field *f;char fname[PATH_MAX ]; -int not_first, count;FILE *ofp;sprintf (fname,"%s.py" , outfile);
@@ -1735,7 +1736,11 @@ static int python_generate_script(struct tep_handle *pevent, const char *outfilefprintf (ofp,"def trace_end():\n" );fprintf (ofp,"\tprint(\"in trace_end\")\n\n" ); -while ((event =trace_find_next_event (pevent, event))) { + nr_events =tep_get_events_count (pevent); + all_events =tep_list_events (pevent,TEP_EVENT_SORT_ID ); + +for (i =0 ; all_events && i < nr_events; i++) { + event = all_events[i];fprintf (ofp,"def %s__%s(" , event->system, event->name);fprintf (ofp,"event_name, " );fprintf (ofp,"context, " );
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 35/57 perf tools: Remove unused trace_find_next_event()
In reply to Arnaldo Carvalho de MeloFrom: "Steven Rostedt (VMware)" <[email protected]>
trace_find_next_event() was buggy and pretty much a useless helper. As there are no more users, just remove it.
--- tools/perf/util/trace-event-parse.c | 31 ----------------------------- tools/perf/util/trace-event.h | 2 -- 2 files changed, 33 deletions(-)
tools/perf/util/trace-event-parse.c +0 -31
@@ -173,37 +173,6 @@ int parse_event_file(struct tep_handle *pevent,return tep_parse_event (pevent, buf, size, sys); } -struct tep_event *trace_find_next_event (struct tep_handle *pevent, -struct tep_event *event) -{ -static int idx; -int events_count; -struct tep_event *all_events; - - all_events =tep_get_first_event (pevent); - events_count =tep_get_events_count (pevent); -if (!pevent || !all_events || events_count <1 ) -return NULL ; - -if (!event) { - idx =0 ; -return all_events; - } - -if (idx < events_count && event == (all_events + idx)) { - idx++; -if (idx == events_count) -return NULL ; -return (all_events + idx); - } - -for (idx =1 ; idx < events_count; idx++) { -if (event == (all_events + (idx -1 ))) -return (all_events + idx); - } -return NULL ; -} -struct flag {const char *name;unsigned long long value;
tools/perf/util/trace-event.h +0 -2
@@ -47,8 +47,6 @@ void parse_saved_cmdline(struct tep_handle *pevent, char *file, unsigned int sizssize_t trace_report (int fd,struct trace_event *tevent,bool repipe); -struct tep_event *trace_find_next_event (struct tep_handle *pevent, -struct tep_event *event);unsigned long long read_size (struct tep_event *event,void *ptr,int size);unsigned long long eval_flag (const char *flag);
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 36/57 libbeauty: Introduce syscall_arg__strtoul_strarray()
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
To go from strarrays strings to its indexes.
--- tools/perf/builtin-trace.c | 6 ++++++ tools/perf/trace/beauty/beauty.h | 3 +++ 2 files changed, 9 insertions(+)
tools/perf/builtin-trace.c +6 -0
@@ -535,6 +535,11 @@ static size_t syscall_arg__scnprintf_strarray(char *bf, size_t size,#define SCA_STRARRAY syscall_arg__scnprintf_strarray +bool syscall_arg__strtoul_strarray (char *bf,size_t size,struct syscall_arg *arg,u64 *ret) +{ +return strarray__strtoul (arg->parm, bf, size, ret); +} +size_t syscall_arg__scnprintf_strarray_flags (char *bf,size_t size,struct syscall_arg *arg) {return strarray__scnprintf_flags (arg->parm, bf, size, arg->show_string_prefix, arg->val);
@@ -824,6 +829,7 @@ static size_t syscall_arg__scnprintf_getrandom_flags(char *bf, size_t size,#define STRARRAY (name, array) \ { .scnprintf = SCA_STRARRAY, \ + .strtoul = STUL_STRARRAY, \ .parm = &strarray__##array, }#define STRARRAY_FLAGS (name, array) \
tools/perf/trace/beauty/beauty.h +3 -0
@@ -122,6 +122,9 @@ unsigned long syscall_arg__val(struct syscall_arg *arg, u8 idx);size_t syscall_arg__scnprintf_strarray_flags (char *bf,size_t size,struct syscall_arg *arg);#define SCA_STRARRAY_FLAGS syscall_arg__scnprintf_strarray_flags +bool syscall_arg__strtoul_strarray (char *bf,size_t size,struct syscall_arg *arg,u64 *ret); +#define STUL_STRARRAY syscall_arg__strtoul_strarray +size_t syscall_arg__scnprintf_x86_irq_vectors (char *bf,size_t size,struct syscall_arg *arg);#define SCA_X86_IRQ_VECTORS syscall_arg__scnprintf_x86_irq_vectors
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 37/57 perf trace: Honour --max-events in processing syscalls:sys_enter_*
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
We were doing this only at the sys_exit syscall tracepoint, as for strace-like we count the pair of sys_enter and sys_exit as one event, but when asking specifically for a the syscalls:sys_enter_NAME tracepoint we need to count each of those as an event.
I.e. things like:
# perf trace --max-events=4 -e syscalls:sys_enter_lseek
0.000 pool/2242 syscalls:sys_enter_lseek(fd: 14<anon_inode:[timerfd]>, offset: 0, whence: CUR)
0.034 pool/2242 syscalls:sys_enter_lseek(fd: 15<anon_inode:[timerfd]>, offset: 0, whence: CUR)
0.051 pool/2242 syscalls:sys_enter_lseek(fd: 16<anon_inode:[timerfd]>, offset: 0, whence: CUR)
2307.900 sshd/30800 syscalls:sys_enter_lseek(fd: 3</usr/lib64/libsystemd.so.0.25.0>, offset: 9032, whence: SET)
#Were going on forever, since we only had sys_enter events.
--- tools/perf/builtin-trace.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-)
tools/perf/builtin-trace.c +7 -6
@@ -2727,12 +2727,6 @@ static int trace__event_handler(struct trace *trace, struct evsel *evsel, }else {trace__fprintf_tp_fields (trace, evsel, sample, thread,NULL ,0 ); } - ++trace->nr_events_printed; - -if (evsel->max_events !=ULONG_MAX && ++evsel->nr_events_printed == evsel->max_events) { -evsel__disable (evsel); -evsel__close (evsel); - } } }
@@ -2743,6 +2737,13 @@ static int trace__event_handler(struct trace *trace, struct evsel *evsel,trace__fprintf_callchain (trace, sample); else if (callchain_ret < 0) pr_err("Problem processing %s callchain, skipping...\n", perf_evsel__name(evsel)); + + ++trace->nr_events_printed; + +if (evsel->max_events !=ULONG_MAX && ++evsel->nr_events_printed == evsel->max_events) { +evsel__disable (evsel); +evsel__close (evsel); + } out:thread__put (thread);return 0 ;
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 38/57 perf trace: Pass a syscall_arg to syscall_arg_fmt->strtoul()
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
With just what we need for the STUL_STRARRAY, i.e. the 'struct strarray' pointer to be used, just like with syscall_arg_fmt->scnprintf() for the other direction (number -> string).
With this all the strarrays that are associated with syscalls can be used with '-e syscalls:sys_enter_SYSCALLNAME --filter', and soon will be possible as well to use with the strace-like shorter form, with just the syscall names, i.e. something like:
-e lseek/whence==END/
For now we have to use the longer form:
# perf trace -e syscalls:sys_enter_lseek
0.000 pool/2242 syscalls:sys_enter_lseek(fd: 14<anon_inode:[timerfd]>, offset: 0, whence: CUR)
0.031 pool/2242 syscalls:sys_enter_lseek(fd: 15<anon_inode:[timerfd]>, offset: 0, whence: CUR)
0.046 pool/2242 syscalls:sys_enter_lseek(fd: 16<anon_inode:[timerfd]>, offset: 0, whence: CUR)
5003.528 pool/2242 syscalls:sys_enter_lseek(fd: 14<anon_inode:[timerfd]>, offset: 0, whence: CUR)
5003.575 pool/2242 syscalls:sys_enter_lseek(fd: 15<anon_inode:[timerfd]>, offset: 0, whence: CUR)
5003.593 pool/2242 syscalls:sys_enter_lseek(fd: 16<anon_inode:[timerfd]>, offset: 0, whence: CUR)
10002.017 pool/2242 syscalls:sys_enter_lseek(fd: 14<anon_inode:[timerfd]>, offset: 0, whence: CUR)
10002.051 pool/2242 syscalls:sys_enter_lseek(fd: 15<anon_inode:[timerfd]>, offset: 0, whence: CUR)
10002.068 pool/2242 syscalls:sys_enter_lseek(fd: 16<anon_inode:[timerfd]>, offset: 0, whence: CUR)
^C# perf trace -e syscalls:sys_enter_lseek --filter="whence!=CUR"
0.000 sshd/24476 syscalls:sys_enter_lseek(fd: 3, offset: 9032, whence: SET)
0.060 sshd/24476 syscalls:sys_enter_lseek(fd: 3</usr/lib64/libcrypt.so.2.0.0>, offset: 9032, whence: SET)
0.187 sshd/24476 syscalls:sys_enter_lseek(fd: 3</usr/lib64/libcrypt.so.2.0.0>, offset: 118632, whence: SET)
0.203 sshd/24476 syscalls:sys_enter_lseek(fd: 3</usr/lib64/libcrypt.so.2.0.0>, offset: 118632, whence: SET)
0.349 sshd/24476 syscalls:sys_enter_lseek(fd: 3</usr/lib64/libcrypt.so.2.0.0>, offset: 61936, whence: SET)
^C#And for those curious about what are those lseek(DSO, offset, SET), well, its the loader:
# perf trace -e syscalls:sys_enter_lseek/max-stack=16/ --filter="whence!=CUR"
0.000 sshd/24495 syscalls:sys_enter_lseek(fd: 3</usr/lib64/libgcrypt.so.20.2.5>, offset: 9032, whence: SET)
__libc_lseek64 (/usr/lib64/ld-2.29.so)
_dl_map_object (/usr/lib64/ld-2.29.so)
0.067 sshd/24495 syscalls:sys_enter_lseek(fd: 3</usr/lib64/libgcrypt.so.20.2.5>, offset: 9032, whence: SET)
__libc_lseek64 (/usr/lib64/ld-2.29.so)
_dl_map_object_from_fd (/usr/lib64/ld-2.29.so)
_dl_map_object (/usr/lib64/ld-2.29.so)
0.198 sshd/24495 syscalls:sys_enter_lseek(fd: 3</usr/lib64/libgcrypt.so.20.2.5>, offset: 118632, whence: SET)
__libc_lseek64 (/usr/lib64/ld-2.29.so)
_dl_map_object (/usr/lib64/ld-2.29.so)
0.219 sshd/24495 syscalls:sys_enter_lseek(fd: 3</usr/lib64/libgcrypt.so.20.2.5>, offset: 118632, whence: SET)
__libc_lseek64 (/usr/lib64/ld-2.29.so)
_dl_map_object_from_fd (/usr/lib64/ld-2.29.so)
_dl_map_object (/usr/lib64/ld-2.29.so)
^C#:-)
With this we can use strings in strarrays in filters, which allows us to reuse all these that are in place for syscalls:
$ find tools/perf/trace/beauty/ -name "*.c" | xargs grep -w DEFINE_STRARRAY tools/perf/trace/beauty/fcntl.c: static DEFINE_STRARRAY(fcntl_setlease, "F_"); tools/perf/trace/beauty/mmap.c: static DEFINE_STRARRAY(mmap_flags, "MAP_"); tools/perf/trace/beauty/mmap.c: static DEFINE_STRARRAY(madvise_advices, "MADV_"); tools/perf/trace/beauty/sync_file_range.c: static DEFINE_STRARRAY(sync_file_range_flags, "SYNC_FILE_RANGE_"); tools/perf/trace/beauty/socket.c: static DEFINE_STRARRAY(socket_ipproto, "IPPROTO_"); tools/perf/trace/beauty/mount_flags.c: static DEFINE_STRARRAY(mount_flags, "MS_"); tools/perf/trace/beauty/pkey_alloc.c: static DEFINE_STRARRAY(pkey_alloc_access_rights, "PKEY_"); tools/perf/trace/beauty/sockaddr.c:DEFINE_STRARRAY(socket_families, "PF_"); tools/perf/trace/beauty/tracepoints/x86_irq_vectors.c:static DEFINE_STRARRAY(x86_irq_vectors, "_VECTOR"); tools/perf/trace/beauty/tracepoints/x86_msr.c:static DEFINE_STRARRAY(x86_MSRs, "MSR_"); tools/perf/trace/beauty/prctl.c: static DEFINE_STRARRAY(prctl_options, "PR_"); tools/perf/trace/beauty/prctl.c: static DEFINE_STRARRAY(prctl_set_mm_options, "PR_SET_MM_"); tools/perf/trace/beauty/fspick.c: static DEFINE_STRARRAY(fspick_flags, "FSPICK_"); tools/perf/trace/beauty/ioctl.c: static DEFINE_STRARRAY(ioctl_tty_cmd, ""); tools/perf/trace/beauty/ioctl.c: static DEFINE_STRARRAY(drm_ioctl_cmds, ""); tools/perf/trace/beauty/ioctl.c: static DEFINE_STRARRAY(sndrv_pcm_ioctl_cmds, ""); tools/perf/trace/beauty/ioctl.c: static DEFINE_STRARRAY(sndrv_ctl_ioctl_cmds, ""); tools/perf/trace/beauty/ioctl.c: static DEFINE_STRARRAY(kvm_ioctl_cmds, ""); tools/perf/trace/beauty/ioctl.c: static DEFINE_STRARRAY(vhost_virtio_ioctl_cmds, ""); tools/perf/trace/beauty/ioctl.c: static DEFINE_STRARRAY(vhost_virtio_ioctl_read_cmds, ""); tools/perf/trace/beauty/ioctl.c: static DEFINE_STRARRAY(perf_ioctl_cmds, ""); tools/perf/trace/beauty/ioctl.c: static DEFINE_STRARRAY(usbdevfs_ioctl_cmds, ""); tools/perf/trace/beauty/fsmount.c: static DEFINE_STRARRAY(fsmount_attr_flags, "MOUNT_ATTR_"); tools/perf/trace/beauty/renameat.c: static DEFINE_STRARRAY(rename_flags, "RENAME_"); tools/perf/trace/beauty/kcmp.c: static DEFINE_STRARRAY(kcmp_types, "KCMP_"); tools/perf/trace/beauty/move_mount.c: static DEFINE_STRARRAY(move_mount_flags, "MOVE_MOUNT_"); $
Well, some, as the mmap flags are like:
$ tools/perf/trace/beauty/mmap_flags.sh
static const char *mmap_flags[] = {
[ilog2(0x40) + 1] = "32BIT",
[ilog2(0x01) + 1] = "SHARED",
[ilog2(0x02) + 1] = "PRIVATE",
[ilog2(0x10) + 1] = "FIXED",
[ilog2(0x20) + 1] = "ANONYMOUS",
[ilog2(0x008000) + 1] = "POPULATE",
[ilog2(0x010000) + 1] = "NONBLOCK",
[ilog2(0x020000) + 1] = "STACK",
[ilog2(0x040000) + 1] = "HUGETLB",
[ilog2(0x080000) + 1] = "SYNC",
[ilog2(0x100000) + 1] = "FIXED_NOREPLACE",
[ilog2(0x0100) + 1] = "GROWSDOWN",
[ilog2(0x0800) + 1] = "DENYWRITE",
[ilog2(0x1000) + 1] = "EXECUTABLE",
[ilog2(0x2000) + 1] = "LOCKED",
[ilog2(0x4000) + 1] = "NORESERVE",
};
$So we'll need a strarray__strtoul_flags() that will break donw the flags into tokens separated by '|' before doing the lookup and then go on reconstructing the value from, say:
# perf trace -e syscalls:sys_enter_mmap --filter="flags==PRIVATE|FIXED|DENYWRITE"
into:
# perf trace -e syscalls:sys_enter_mmap --filter="flags==0x2|0x10|0x0800"
and finally into:
# perf trace -e syscalls:sys_enter_mmap --filter="flags==0x812"
That is what we see if we don't use the augmented view obtained from:
# perf trace -e mmap <SNIP> 211792.885 procmail/15393 mmap(addr: 0x7fcd11645000, len: 8192, prot: READ, flags: PRIVATE|FIXED|DENYWRITE, fd: 8, off: 0xa000) = 0x7fcd11645000 <SNIP>
But plain use tracefs:
procmail-15559 [000] .... 54557.178262: sys_mmap(addr: 7f5c9bf7a000, len: 9b000, prot: 1, flags: 812, fd: 3, off: a9000)
--- tools/perf/builtin-trace.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)
tools/perf/builtin-trace.c +5 -1
@@ -3696,7 +3696,11 @@ static int trace__expand_filter(struct trace *trace __maybe_unused, struct evselif (fmt->strtoul) {u64 val; -if (fmt->strtoul (right, right_size,NULL , &val)) { +struct syscall_arg syscall_arg = { + .parm = fmt->parm, + }; + +if (fmt->strtoul (right, right_size, &syscall_arg, &val)) {char *n, expansion[19 ];int expansion_lenght =scnprintf (expansion,sizeof (expansion),"%#" PRIx64, val);int expansion_offset = right - new_filter;
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 39/57 perf list: Hide deprecated events by default
In reply to Arnaldo Carvalho de MeloFrom: Jin Yao <[email protected]>
There are some deprecated events listed by perf list. But we can't remove them from perf list with ease because some old scripts may use them.
Deprecated events are old names of renamed events. When an event gets renamed the old name is kept around for some time and marked with Deprecated. The newer Intel event lists in the tree already have these headers.
So we need to keep them in the event list, but provide a new option to show them. The new option is "--deprecated".
With this patch, the deprecated events are hidden by default but they can be displayed when option "--deprecated" is enabled.
--- tools/perf/Documentation/perf-list.txt | 3 +++ tools/perf/builtin-list.c | 14 ++++++++++---- tools/perf/pmu-events/jevents.c | 26 ++++++++++++++++++++------ tools/perf/pmu-events/jevents.h | 3 ++- tools/perf/pmu-events/pmu-events.h | 1 + tools/perf/util/parse-events.c | 4 ++-- tools/perf/util/parse-events.h | 2 +- tools/perf/util/pmu.c | 17 +++++++++++++---- tools/perf/util/pmu.h | 4 +++- 9 files changed, 55 insertions(+), 19 deletions(-)
tools/perf/Documentation/perf-list.txt +3 -0
@@ -36,6 +36,9 @@ Enable debugging output. Print how named events are resolved internally into perf events, and also any extra expressions computed by perf stat. +--deprecated:: +Print deprecated events. By default the deprecated events are hidden. + [[EVENT_MODIFIERS]] EVENT MODIFIERS ---------------
tools/perf/builtin-list.c +10 -4
@@ -26,6 +26,7 @@ int cmd_list(int argc, const char **argv)int i;bool raw_dump = false;bool long_desc_flag = false; +bool deprecated = false;struct option list_options[] = {OPT_BOOLEAN (0 ,"raw-dump" , &raw_dump,"Dump raw events" ), OPT_BOOLEAN('d', "desc", &desc_flag,
@@ -34,6 +35,8 @@ int cmd_list(int argc, const char **argv) "Print longer event descriptions."),OPT_BOOLEAN (0 ,"details" , &details_flag,"Print information on the perf event names and expressions used internally by events." ), +OPT_BOOLEAN (0 ,"deprecated" , &deprecated, +"Print deprecated events." ),OPT_INCR (0 ,"debug" , &verbose,"Enable debugging output" ),OPT_END ()
@@ -55,7 +58,7 @@ int cmd_list(int argc, const char **argv)if (argc ==0 ) {print_events (NULL , raw_dump, !desc_flag, long_desc_flag, - details_flag); + details_flag, deprecated);return 0 ; }
@@ -78,7 +81,8 @@ int cmd_list(int argc, const char **argv)print_hwcache_events (NULL , raw_dump); else if (strcmp(argv[i], "pmu") == 0) print_pmu_events(NULL, raw_dump, !desc_flag, - long_desc_flag, details_flag); + long_desc_flag, details_flag, + deprecated); else if (strcmp(argv[i], "sdt") == 0) print_sdt_events(NULL, NULL, raw_dump); else if (strcmp(argv[i], "metric") == 0 || strcmp(argv[i], "metrics") == 0)
@@ -91,7 +95,8 @@ int cmd_list(int argc, const char **argv)if (sep ==NULL ) {print_events (argv[i], raw_dump, !desc_flag, long_desc_flag, - details_flag); + details_flag, + deprecated);continue ; } sep_idx = sep - argv[i];
@@ -117,7 +122,8 @@ int cmd_list(int argc, const char **argv)print_hwcache_events (s, raw_dump);print_pmu_events (s, raw_dump, !desc_flag, long_desc_flag, - details_flag); + details_flag, + deprecated);print_tracepoint_events (NULL , s, raw_dump);print_sdt_events (NULL , s, raw_dump);metricgroup__print (true, true, s, raw_dump, details_flag);
tools/perf/pmu-events/jevents.c +20 -6
@@ -322,7 +322,8 @@ static int print_events_table_entry(void *data, char *name, char *event, char *desc, char *long_desc, char *pmu, char *unit, char *perpkg, char *metric_expr, - char *metric_name, char *metric_group) + char *metric_name, char *metric_group, + char *deprecated) { struct perf_entry_data *pd = data;FILE *outfp = pd->outfp;
@@ -354,6 +355,8 @@ static int print_events_table_entry(void *data, char *name, char *event,fprintf (outfp,"\t.metric_name = \"%s\",\n" , metric_name);if (metric_group)fprintf (outfp,"\t.metric_group = \"%s\",\n" , metric_group); +if (deprecated) +fprintf (outfp,"\t.deprecated = \"%s\",\n" , deprecated);fprintf (outfp,"},\n" );return 0 ;
@@ -371,6 +374,7 @@ struct event_struct {char *metric_expr;char *metric_name;char *metric_group; +char *deprecated; };#define ADD_EVENT_FIELD (field) do { if (field) { \
@@ -398,6 +402,7 @@ struct event_struct {op (metric_expr); \op (metric_name); \op (metric_group); \ +op (deprecated); \ }while (0 ) static LIST_HEAD(arch_std_events);
@@ -416,7 +421,8 @@ static void free_arch_std_events(void)static int save_arch_std_events (void *data,char *name,char *event,char *desc,char *long_desc,char *pmu,char *unit,char *perpkg,char *metric_expr, -char *metric_name,char *metric_group) +char *metric_name,char *metric_group, +char *deprecated) {struct event_struct *es;
@@ -479,7 +485,8 @@ static int try_fixup(const char *fn, char *arch_std, char **event, char **desc, char **name, char **long_desc, char **pmu, char **filter, char **perpkg, char **unit, char **metric_expr, char **metric_name, - char **metric_group, unsigned long long eventcode) + char **metric_group, unsigned long long eventcode, + char **deprecated) { /* try to find matching event from arch standard values */ struct event_struct *es;
@@ -507,7 +514,8 @@ int json_events(const char *fn,char *long_desc, char *pmu, char *unit, char *perpkg, char *metric_expr, - char *metric_name, char *metric_group), + char *metric_name, char *metric_group, + char *deprecated), void *data) { int err;
@@ -536,6 +544,7 @@ int json_events(const char *fn,char *metric_expr =NULL ;char *metric_name =NULL ;char *metric_group =NULL ; +char *deprecated =NULL ;char *arch_std =NULL ;unsigned long long eventcode =0 ;struct msrmap *msr =NULL ;
@@ -614,6 +623,8 @@ int json_events(const char *fn,addfield (map, &unit,"" ,"" , val); }else if (json_streq (map, field,"PerPkg" )) {addfield (map, &perpkg,"" ,"" , val); + }else if (json_streq (map, field,"Deprecated" )) { +addfield (map, &deprecated,"" ,"" , val); }else if (json_streq (map, field,"MetricName" )) {addfield (map, &metric_name,"" ,"" , val); }else if (json_streq (map, field,"MetricGroup" )) {
@@ -658,12 +669,14 @@ int json_events(const char *fn, err =try_fixup (fn, arch_std, &event, &desc, &name, &long_desc, &pmu, &filter, &perpkg, &unit, &metric_expr, &metric_name, - &metric_group, eventcode); + &metric_group, eventcode, + &deprecated);if (err) goto free_strings; } err =func (data, name,real_event (name, event), desc, long_desc, - pmu, unit, perpkg, metric_expr, metric_name, metric_group); + pmu, unit, perpkg, metric_expr, metric_name, + metric_group, deprecated); free_strings:free (event);free (desc);
@@ -673,6 +686,7 @@ int json_events(const char *fn,free (pmu);free (filter);free (perpkg); +free (deprecated);free (unit);free (metric_expr);free (metric_name);
tools/perf/pmu-events/jevents.h +2 -1
@@ -7,7 +7,8 @@ int json_events(const char *fn,char *long_desc, char *pmu, char *unit, char *perpkg, char *metric_expr, - char *metric_name, char *metric_group), + char *metric_name, char *metric_group, + char *deprecated), void *data);char *get_cpu_str (void );
tools/perf/pmu-events/pmu-events.h +1 -0
@@ -17,6 +17,7 @@ struct pmu_event {const char *metric_expr;const char *metric_name;const char *metric_group; +const char *deprecated; };/*
tools/perf/util/parse-events.c +2 -2
@@ -2600,7 +2600,7 @@ void print_symbol_events(const char *event_glob, unsigned type,* Print the help text for the event symbols: */ void print_events (const char *event_glob,bool name_only,bool quiet_flag, -bool long_desc,bool details_flag) +bool long_desc,bool details_flag,bool deprecated) {print_symbol_events (event_glob,PERF_TYPE_HARDWARE , event_symbols_hw,PERF_COUNT_HW_MAX , name_only);
@@ -2612,7 +2612,7 @@ void print_events(const char *event_glob, bool name_only, bool quiet_flag,print_hwcache_events (event_glob, name_only);print_pmu_events (event_glob, name_only, quiet_flag, long_desc, - details_flag); + details_flag, deprecated);if (event_glob !=NULL )return ;
tools/perf/util/parse-events.h +1 -1
@@ -195,7 +195,7 @@ void parse_events_evlist_error(struct parse_events_state *parse_state, int idx, const char *str);void print_events (const char *event_glob,bool name_only,bool quiet, -bool long_desc,bool details_flag); +bool long_desc,bool details_flag,bool deprecated);struct event_symbol {const char *symbol;
tools/perf/util/pmu.c +13 -4
@@ -308,7 +308,8 @@ static int __perf_pmu__new_alias(struct list_head *list, char *dir, char *name, char *long_desc, char *topic, char *unit, char *perpkg, char *metric_expr, - char *metric_name) + char *metric_name, + char *deprecated) { struct parse_events_term *term;struct perf_pmu_alias *alias;
@@ -325,6 +326,7 @@ static int __perf_pmu__new_alias(struct list_head *list, char *dir, char *name, alias->unit[0 ] ='\0' ; alias->per_pkg = false; alias->snapshot = false; + alias->deprecated = false; ret =parse_events_terms (&alias->terms, val);if (ret) {
@@ -379,6 +381,9 @@ static int __perf_pmu__new_alias(struct list_head *list, char *dir, char *name, alias->per_pkg = perpkg &&sscanf (perpkg,"%d" , &num) ==1 && num ==1 ; alias->str =strdup (newval); +if (deprecated) + alias->deprecated = true; +if (!perf_pmu_merge_alias (alias, list))list_add_tail (&alias->list, list);
@@ -400,7 +405,7 @@ static int perf_pmu__new_alias(struct list_head *list, char *dir, char *name, FIstrim (buf);return __perf_pmu__new_alias (list, dir, name,NULL , buf,NULL ,NULL ,NULL , -NULL ,NULL ,NULL ); +NULL ,NULL ,NULL ,NULL ); }static inline bool pmu_alias_info_file (char *name)
@@ -787,7 +792,8 @@ static void pmu_add_cpu_aliases(struct list_head *head, struct perf_pmu *pmu) (char *)pe->long_desc, (char *)pe->topic, (char *)pe->unit, (char *)pe->perpkg, (char *)pe->metric_expr, - (char *)pe->metric_name); + (char *)pe->metric_name, + (char *)pe->deprecated); } }
@@ -1383,7 +1389,7 @@ static void wordwrap(char *s, int start, int max, int corr) }void print_pmu_events (const char *event_glob,bool name_only,bool quiet_flag, -bool long_desc,bool details_flag) +bool long_desc,bool details_flag,bool deprecated) {struct perf_pmu *pmu;struct perf_pmu_alias *alias;
@@ -1414,6 +1420,9 @@ void print_pmu_events(const char *event_glob, bool name_only, bool quiet_flag,format_alias (buf,sizeof (buf), pmu, alias);bool is_cpu = !strcmp (pmu->name,"cpu" ); +if (alias->deprecated && !deprecated) +continue ; + if (event_glob != NULL && !(strglobmatch_nocase(name, event_glob) || (!is_cpu && strglobmatch_nocase(alias->name,
tools/perf/util/pmu.h +3 -1
@@ -57,6 +57,7 @@ struct perf_pmu_alias {double scale;bool per_pkg;bool snapshot; +bool deprecated;char *metric_expr;char *metric_name; };
@@ -85,7 +86,8 @@ int perf_pmu__format_parse(char *dir, struct list_head *head);struct perf_pmu *perf_pmu__scan (struct perf_pmu *pmu);void print_pmu_events (const char *event_glob,bool name_only,bool quiet, -bool long_desc,bool details_flag); +bool long_desc,bool details_flag, +bool deprecated);bool pmu_have_event (const char *pname,const char *name); int perf_pmu__scan_file(struct perf_pmu *pmu, const char *name, const char *fmt, ...) __scanf(3, 4);
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 40/57 perf tests: Remove needless headers for bp_account
In reply to Arnaldo Carvalho de MeloFrom: Leo Yan <[email protected]>
A few headers are not needed and were introduced by copying from other test file. This patch removes the needless headers for the breakpoint accounting testing.
--- tools/perf/tests/bp_account.c | 4 ---- 1 file changed, 4 deletions(-)
tools/perf/tests/bp_account.c +0 -4
@@ -10,11 +10,7 @@#include <unistd.h> #include <string.h> #include <sys/ioctl.h> -#include <time.h> #include <fcntl.h> -#include <signal.h> -#include <sys/mman.h> -#include <linux/compiler.h> #include <linux/hw_breakpoint.h> #include "tests.h"
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 41/57 perf tests bp_account: Add dedicated checking helper is_supported()
In reply to Arnaldo Carvalho de MeloFrom: Leo Yan <[email protected]>
The arm architecture supports breakpoint accounting but it doesn't support breakpoint overflow signal handling. The current code uses the same checking helper, thus it disables both testings (bp_account and bp_signal) for arm platform.
For handling two testings separately, this patch adds a dedicated checking helper is_supported() for breakpoint accounting testing, thus it allows supporting breakpoint accounting testing on arm platform; the old helper test__bp_signal_is_supported() is only used to checking for breakpoint overflow signal testing.
--- tools/perf/tests/bp_account.c | 16 ++++++++++++++++ tools/perf/tests/builtin-test.c | 2 +- tools/perf/tests/tests.h | 1 + 3 files changed, 18 insertions(+), 1 deletion(-)
tools/perf/tests/bp_account.c +16 -0
@@ -188,3 +188,19 @@ int test__bp_accounting(struct test *test __maybe_unused, int subtest __maybe_unreturn bp_accounting (wp_cnt, share); } + +bool test__bp_account_is_supported (void ) +{ +/* +* PowerPC and S390 do not support creation of instruction +* breakpoints using the perf_event interface. +* +* Just disable the test for these architectures until these +* issues are resolved. +*/ +#if defined(__powerpc__) || defined(__s390x__) +return false; +#else +return true; +#endif +}
tools/perf/tests/builtin-test.c +1 -1
@@ -121,7 +121,7 @@ static struct test generic_tests[] = { { .desc = "Breakpoint accounting", .func = test__bp_accounting, - .is_supported = test__bp_signal_is_supported, + .is_supported = test__bp_account_is_supported, }, { .desc = "Watchpoint",
tools/perf/tests/tests.h +1 -0
@@ -111,6 +111,7 @@ int test__map_groups__merge_in(struct test *t, int subtest);int test__time_utils (struct test *t,int subtest);bool test__bp_signal_is_supported (void ); +bool test__bp_account_is_supported (void );bool test__wp_is_supported (void );#if defined(__arm__) || defined(__aarch64__)
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 42/57 perf tests: Disable bp_signal testing for arm64
In reply to Arnaldo Carvalho de MeloFrom: Leo Yan <[email protected]>
As there are several discussions for enabling perf breakpoint signal testing on arm64 platform: arm64 needs to rely on single-step to execute the breakpointed instruction and then reinstall the breakpoint exception handler. But if we hook the breakpoint with a signal, the signal handler will do the stepping rather than the breakpointed instruction, this causes infinite loops as below:
Kernel space | Userspace
---------------------------------|--------------------------------
| __test_function() -> hit
| breakpoint
breakpoint_handler() |
`-> user_enable_single_step() |
do_signal() |
| sig_handler() -> Step one
| instruction and
| trap to kernel
single_step_handler() |
`-> reinstall_suspended_bps() |
| __test_function() -> hit
| breakpoint again and
| repeat up flow infinitelyAs Will Deacon mentioned [1]: "that we require the overflow handler to do the stepping on arm/arm64, which is relied upon by GDB/ptrace. The hw_breakpoint code is a complete disaster so my preference would be to rip out the perf part and just implement something directly in ptrace, but it's a pretty horrible job". Though Will commented this on arm architecture, but the comment also can apply on arm64 architecture.
For complete information, I searched online and found a few years back, Wang Nan sent one patch 'arm64: Store breakpoint single step state into pstate' [2]; the patch tried to resolve this issue by avoiding single stepping in signal handler and defer to enable the signal stepping when return to __test_function(). The fixing was not merged due to the concern for missing to handle different usage cases.
Based on the info, the most feasible way is to skip Perf breakpoint signal testing for arm64 and this could avoid the duplicate investigation efforts when people see the failure. This patch skips this case on arm64 platform, which is same with arm architecture.
[1] https://lkml.org/lkml/2018/11/15/205
[2] https://lkml.org/lkml/2015/12/23/477
--- tools/perf/tests/bp_signal.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-)
tools/perf/tests/bp_signal.c +6 -9
@@ -49,14 +49,6 @@ asm ("__test_function:\n" "incq (%rdi)\n" "ret\n"); -#elif defined (__aarch64__) -extern void __test_function (volatile long *ptr); -asm ( -".globl __test_function\n" -"__test_function:\n" -"str x30, [x0]\n" -"ret\n" ); -#else static void __test_function (volatile long *ptr) {
@@ -302,10 +294,15 @@ bool test__bp_signal_is_supported(void)* stepping into the SIGIO handler and getting stuck on the * breakpointed instruction. * +* Since arm64 has the same issue with arm for the single-step +* handling, this case also gets suck on the breakpointed +* instruction. +* * Just disable the test for these architectures until these * issues are resolved. */ -#if defined(__powerpc__) || defined(__s390x__) || defined(__arm__) +#if defined(__powerpc__) || defined(__s390x__) || defined(__arm__) || \ + defined(__aarch64__)return false;#else return true;
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 43/57 libperf: Introduce perf_evlist__for_each_mmap()
In reply to Arnaldo Carvalho de MeloFrom: Jiri Olsa <[email protected]>
Add the perf_evlist__for_each_mmap() function and export it in the perf/evlist.h header, so that the user can iterate through 'struct perf_mmap' objects.
Add a internal perf_mmap__link() function to do the actual linking.
--- tools/perf/lib/evlist.c | 26 +++++++++++++++++++++++- tools/perf/lib/include/internal/evlist.h | 2 ++ tools/perf/lib/include/internal/mmap.h | 5 +++-- tools/perf/lib/include/perf/evlist.h | 9 ++++++++ tools/perf/lib/libperf.map | 1 + tools/perf/lib/mmap.c | 6 ++++-- tools/perf/util/evlist.c | 4 +++- 7 files changed, 47 insertions(+), 6 deletions(-)
tools/perf/lib/evlist.c +25 -1
@@ -347,6 +347,8 @@ static struct perf_mmap* perf_evlist__alloc_mmap(struct perf_evlist *evlist, booreturn NULL ;for (i =0 ; i < evlist->nr_mmaps; i++) { +struct perf_mmap *prev = i ? &map[i -1 ] :NULL ; +/* * When the perf_mmap() call is made we grab one refcount, plus * one extra to let perf_mmap__consume() get the last
@@ -356,7 +358,7 @@ static struct perf_mmap* perf_evlist__alloc_mmap(struct perf_evlist *evlist, boo* Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and * thus does perf_mmap__get() on it. */ -perf_mmap__init (&map[i], overwrite,NULL ); +perf_mmap__init (&map[i], prev, overwrite,NULL ); }return map;
@@ -405,6 +407,15 @@ perf_evlist__mmap_cb_mmap(struct perf_mmap *map, struct perf_mmap_param *mp,return perf_mmap__mmap (map, mp, output, cpu); } +static void perf_evlist__set_mmap_first (struct perf_evlist *evlist,struct perf_mmap *map, +bool overwrite) +{ +if (overwrite) + evlist->mmap_ovw_first = map; +else + evlist->mmap_first = map; +} + static int mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops, int idx, struct perf_mmap_param *mp, int cpu_idx,
@@ -460,6 +471,9 @@ mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,if (ops->mmap (map, mp, *output, evlist_cpu) <0 )return -1 ; + +if (!idx) +perf_evlist__set_mmap_first (evlist, map, overwrite); }else {if (ioctl (fd,PERF_EVENT_IOC_SET_OUTPUT , *output) !=0 )return -1 ;
@@ -605,3 +619,13 @@ void perf_evlist__munmap(struct perf_evlist *evlist)zfree (&evlist->mmap);zfree (&evlist->mmap_ovw); } + +struct perf_mmap * +perf_evlist__next_mmap (struct perf_evlist *evlist,struct perf_mmap *map, +bool overwrite) +{ +if (map) +return map->next; + +return overwrite ? evlist->mmap_ovw_first : evlist->mmap_first; +}
tools/perf/lib/include/internal/evlist.h +2 -0
@@ -25,6 +25,8 @@ struct perf_evlist {struct hlist_head heads[PERF_EVLIST__HLIST_SIZE ];struct perf_mmap *mmap;struct perf_mmap *mmap_ovw; +struct perf_mmap *mmap_first; +struct perf_mmap *mmap_ovw_first; }; typedef void
tools/perf/lib/include/internal/mmap.h +3 -2
@@ -32,6 +32,7 @@ struct perf_mmap {u64 flush;libperf_unmap_cb_t unmap_cb;char event_copy[PERF_SAMPLE_MAX_SIZE ]__aligned(8) ; +struct perf_mmap *next; };struct perf_mmap_param {
@@ -41,8 +42,8 @@ struct perf_mmap_param {size_t perf_mmap__mmap_len (struct perf_mmap *map); -void perf_mmap__init (struct perf_mmap *map,bool overwrite, -libperf_unmap_cb_t unmap_cb); +void perf_mmap__init (struct perf_mmap *map,struct perf_mmap *prev, +bool overwrite,libperf_unmap_cb_t unmap_cb);int perf_mmap__mmap (struct perf_mmap *map,struct perf_mmap_param *mp,int fd,int cpu);void perf_mmap__munmap (struct perf_mmap *map);
tools/perf/lib/include/perf/evlist.h +9 -0
@@ -3,6 +3,7 @@#define __LIBPERF_EVLIST_H#include <perf/core.h> +#include <stdbool.h> struct perf_evlist ;struct perf_evsel ;
@@ -38,4 +39,12 @@ LIBPERF_API int perf_evlist__filter_pollfd(struct perf_evlist *evlist, LIBPERF_API int perf_evlist__mmap(struct perf_evlist *evlist, int pages); LIBPERF_API void perf_evlist__munmap(struct perf_evlist *evlist); +LIBPERF_API struct perf_mmap *perf_evlist__next_mmap(struct perf_evlist *evlist, +struct perf_mmap *map, +bool overwrite); +#define perf_evlist__for_each_mmap (evlist, pos, overwrite) \ + for ((pos) = perf_evlist__next_mmap((evlist), NULL, overwrite); \ + (pos) != NULL; \ + (pos) = perf_evlist__next_mmap((evlist), (pos), overwrite)) +#endif /* __LIBPERF_EVLIST_H */
tools/perf/lib/libperf.map +1 -0
@@ -43,6 +43,7 @@ LIBPERF_0.0.1 { perf_evlist__mmap; perf_evlist__munmap; perf_evlist__filter_pollfd; + perf_evlist__next_mmap; perf_mmap__consume; perf_mmap__read_init; perf_mmap__read_done;
tools/perf/lib/mmap.c +4 -2
@@ -13,13 +13,15 @@#include <linux/kernel.h> #include "internal.h" -void perf_mmap__init (struct perf_mmap *map,bool overwrite, -libperf_unmap_cb_t unmap_cb) +void perf_mmap__init (struct perf_mmap *map,struct perf_mmap *prev, +bool overwrite,libperf_unmap_cb_t unmap_cb) { map->fd =-1 ; map->overwrite = overwrite; map->unmap_cb = unmap_cb;refcount_set (&map->refcnt,0 ); +if (prev) + prev->next = map; }size_t perf_mmap__mmap_len (struct perf_mmap *map)
tools/perf/util/evlist.c +3 -1
@@ -607,6 +607,8 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,return NULL ;for (i =0 ; i < evlist->core.nr_mmaps; i++) { +struct perf_mmap *prev = i ? &map[i -1 ].core :NULL ; +/* * When the perf_mmap() call is made we grab one refcount, plus * one extra to let perf_mmap__consume() get the last
@@ -616,7 +618,7 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,* Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and * thus does perf_mmap__get() on it. */ -perf_mmap__init (&map[i].core, overwrite, perf_mmap__unmap_cb); +perf_mmap__init (&map[i].core, prev, overwrite, perf_mmap__unmap_cb); }return map;
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 44/57 libperf: Move mmap allocation to perf_evlist__mmap_ops::get
In reply to Arnaldo Carvalho de MeloFrom: Jiri Olsa <[email protected]>
Move allocation of the mmap array into perf_evlist__mmap_ops::get, to centralize the mmap allocation.
Also move nr_mmap setup to perf_evlist__mmap_ops so it's centralized and shared by both perf and libperf mmap code.
--- tools/perf/lib/evlist.c | 42 ++++++++++++++++++++++++---------------- tools/perf/util/evlist.c | 24 +++++++++-------------- 2 files changed, 34 insertions(+), 32 deletions(-)
tools/perf/lib/evlist.c +25 -17
@@ -338,10 +338,6 @@ static struct perf_mmap* perf_evlist__alloc_mmap(struct perf_evlist *evlist, booint i;struct perf_mmap *map; - evlist->nr_mmaps =perf_cpu_map__nr (evlist->cpus); -if (perf_cpu_map__empty (evlist->cpus)) - evlist->nr_mmaps =perf_thread_map__nr (evlist->threads); - map =zalloc (evlist->nr_mmaps *sizeof (struct perf_mmap ));if (!map)return NULL ;
@@ -384,18 +380,22 @@ static void perf_evlist__set_sid_idx(struct perf_evlist *evlist,static struct perf_mmap *perf_evlist__mmap_cb_get (struct perf_evlist *evlist,bool overwrite,int idx) { -struct perf_mmap *map = &evlist->mmap[idx]; +struct perf_mmap *maps; -if (overwrite) { -if (!evlist->mmap_ovw) { - evlist->mmap_ovw =perf_evlist__alloc_mmap (evlist, true); -if (!evlist->mmap_ovw) -return NULL ; - } - map = &evlist->mmap_ovw[idx]; + maps = overwrite ? evlist->mmap_ovw : evlist->mmap; + +if (!maps) { + maps =perf_evlist__alloc_mmap (evlist, overwrite); +if (!maps) +return NULL ; + +if (overwrite) + evlist->mmap_ovw = maps; +else + evlist->mmap = maps; } -return map; +return &maps[idx]; }#define FD (e, x, y) (*(int *) xyarray__entry(e->fd, x, y))
@@ -556,6 +556,17 @@ mmap_per_cpu(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,return -1 ; } +static int perf_evlist__nr_mmaps (struct perf_evlist *evlist) +{ +int nr_mmaps; + + nr_mmaps =perf_cpu_map__nr (evlist->cpus); +if (perf_cpu_map__empty (evlist->cpus)) + nr_mmaps =perf_thread_map__nr (evlist->threads); + +return nr_mmaps; +} +int perf_evlist__mmap_ops (struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops,struct perf_mmap_param *mp)
@@ -567,10 +578,7 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,if (!ops || !ops->get || !ops->mmap)return -EINVAL ; -if (!evlist->mmap) - evlist->mmap =perf_evlist__alloc_mmap (evlist, false); -if (!evlist->mmap) -return -ENOMEM ; + evlist->nr_mmaps =perf_evlist__nr_mmaps (evlist);perf_evlist__for_each_entry (evlist, evsel) { if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
tools/perf/util/evlist.c +9 -15
@@ -599,9 +599,6 @@ static struct mmap *evlist__alloc_mmap(struct evlist *evlist,int i;struct mmap *map; - evlist->core.nr_mmaps =perf_cpu_map__nr (evlist->core.cpus); -if (perf_cpu_map__empty (evlist->core.cpus)) - evlist->core.nr_mmaps =perf_thread_map__nr (evlist->core.threads); map =zalloc (evlist->core.nr_mmaps *sizeof (struct mmap ));if (!map)return NULL ;
@@ -639,19 +636,21 @@ static struct perf_mmap* perf_evlist__mmap_cb_get(struct perf_evlist *_evlist, bool overwrite, int idx) { struct evlist *evlist = container_of(_evlist, struct evlist, core); -struct mmap *maps = evlist->mmap; +struct mmap *maps; -if (overwrite) { - maps = evlist->overwrite_mmap; + maps = overwrite ? evlist->overwrite_mmap : evlist->mmap; -if (!maps) { - maps =evlist__alloc_mmap (evlist, true); -if (!maps) -return NULL ; +if (!maps) { + maps =evlist__alloc_mmap (evlist, overwrite); +if (!maps) +return NULL ; +if (overwrite) { evlist->overwrite_mmap = maps;if (evlist->bkw_mmap_state ==BKW_MMAP_NOTREADY )perf_evlist__toggle_bkw_mmap (evlist,BKW_MMAP_RUNNING ); + }else { + evlist->mmap = maps; } }
@@ -812,11 +811,6 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, .mmap = perf_evlist__mmap_cb_mmap, }; -if (!evlist->mmap) - evlist->mmap =evlist__alloc_mmap (evlist, false); -if (!evlist->mmap) -return -ENOMEM ; - evlist->core.mmap_len =evlist__mmap_size (pages);pr_debug ("mmap size %zuB\n" , evlist->core.mmap_len); mp.core.mask = evlist->core.mmap_len - page_size -1 ;
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 45/57 libperf: Move mask setup to perf_evlist__mmap_ops()
In reply to Arnaldo Carvalho de MeloFrom: Jiri Olsa <[email protected]>
Move the mask setup to perf_evlist__mmap_ops(), because it's the same on both perf and libperf path.
--- tools/perf/lib/evlist.c | 3 ++- tools/perf/util/evlist.c | 1 - 2 files changed, 2 insertions(+), 2 deletions(-)
tools/perf/lib/evlist.c +2 -1
@@ -578,6 +578,8 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,if (!ops || !ops->get || !ops->mmap)return -EINVAL ; + mp->mask = evlist->mmap_len - page_size -1 ; + evlist->nr_mmaps =perf_evlist__nr_mmaps (evlist);perf_evlist__for_each_entry (evlist, evsel) {
@@ -605,7 +607,6 @@ int perf_evlist__mmap(struct perf_evlist *evlist, int pages) }; evlist->mmap_len = (pages +1 ) * page_size; - mp.mask = evlist->mmap_len - page_size -1 ;return perf_evlist__mmap_ops (evlist, &ops, &mp); }
tools/perf/util/evlist.c +0 -1
@@ -813,7 +813,6 @@ int evlist__mmap_ex(struct evlist *evlist, unsigned int pages, evlist->core.mmap_len =evlist__mmap_size (pages);pr_debug ("mmap size %zuB\n" , evlist->core.mmap_len); - mp.core.mask = evlist->core.mmap_len - page_size -1 ;auxtrace_mmap_params__init (&mp.auxtrace_mp, evlist->core.mmap_len, auxtrace_pages, auxtrace_overwrite);
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 46/57 libperf: Link static tests with libapi.a
In reply to Arnaldo Carvalho de MeloFrom: Jiri Olsa <[email protected]>
Both static and dynamic tests needs to link with libapi.a, because it's using its functions. Also include path for libapi includes.
--- tools/perf/lib/Makefile | 1 + tools/perf/lib/tests/Makefile | 6 +++--- 2 files changed, 4 insertions(+), 3 deletions(-)
tools/perf/lib/Makefile +1 -0
@@ -107,6 +107,7 @@ else endif LIBAPI = $(API_PATH)libapi.a +export LIBAPI $(LIBAPI): FORCE $(Q)$(MAKE) -C $(LIB_DIR) O=$(OUTPUT) $(OUTPUT)libapi.a
tools/perf/lib/tests/Makefile +3 -3
@@ -16,13 +16,13 @@ all: include $(srctree)/tools/scripts/Makefile.include -INCLUDE = -I$(srctree)/tools/perf/lib/include -I$(srctree)/tools/include +INCLUDE = -I$(srctree)/tools/perf/lib/include -I$(srctree)/tools/include -I$(srctree)/tools/lib $(TESTS_A): FORCE - $(QUIET_LINK)$(CC) $(INCLUDE) $(CFLAGS) -o $@ $(subst -a,.c,$@) ../libperf.a + $(QUIET_LINK)$(CC) $(INCLUDE) $(CFLAGS) -o $@ $(subst -a,.c,$@) ../libperf.a $(LIBAPI) $(TESTS_SO): FORCE - $(QUIET_LINK)$(CC) $(INCLUDE) $(CFLAGS) -L.. -o $@ $(subst -so,.c,$@) -lperf + $(QUIET_LINK)$(CC) $(INCLUDE) $(CFLAGS) -L.. -o $@ $(subst -so,.c,$@) $(LIBAPI) -lperf all: $(TESTS_A) $(TESTS_SO)
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 47/57 libperf: Add tests_mmap_thread test
In reply to Arnaldo Carvalho de MeloFrom: Jiri Olsa <[email protected]>
Add mmaping tests that generates 100 prctl calls in monitored child process and validates it gets 100 events in ring buffer.
Committer tests:
# make -C tools/perf/lib tests
make: Entering directory '/home/acme/git/perf/tools/perf/lib'
LINK test-cpumap-a
LINK test-threadmap-a
LINK test-evlist-a
LINK test-evsel-a
LINK test-cpumap-so
LINK test-threadmap-so
LINK test-evlist-so
LINK test-evsel-so
running static:
- running test-cpumap.c...OK
- running test-threadmap.c...OK
- running test-evlist.c...OK
- running test-evsel.c...OK
running dynamic:
- running test-cpumap.c...OK
- running test-threadmap.c...OK
- running test-evlist.c...OK
- running test-evsel.c...OK
make: Leaving directory '/home/acme/git/perf/tools/perf/lib'
#--- tools/perf/lib/tests/test-evlist.c | 119 +++++++++++++++++++++++++++++ 1 file changed, 119 insertions(+)
tools/perf/lib/tests/test-evlist.c +119 -0
@@ -1,12 +1,21 @@// SPDX-License-Identifier: GPL-2.0 #include <stdio.h> #include <stdarg.h> +#include <unistd.h> +#include <stdlib.h> #include <linux/perf_event.h> +#include <linux/limits.h> +#include <sys/types.h> +#include <sys/wait.h> +#include <sys/prctl.h> #include <perf/cpumap.h> #include <perf/threadmap.h> #include <perf/evlist.h> #include <perf/evsel.h> +#include <perf/mmap.h> +#include <perf/event.h> #include <internal/tests.h> +#include <api/fs/fs.h> static int libperf_print (enum libperf_print_level level,const char *fmt,va_list ap)
@@ -181,6 +190,115 @@ static int test_stat_thread_enable(void)return 0 ; } +static int test_mmap_thread (void ) +{ +struct perf_evlist *evlist; +struct perf_evsel *evsel; +struct perf_mmap *map; +struct perf_cpu_map *cpus; +struct perf_thread_map *threads; +struct perf_event_attr attr = { + .type =PERF_TYPE_TRACEPOINT , + .sample_period =1 , + .wakeup_watermark =1 , + .disabled =1 , + }; +char path[PATH_MAX ]; +int id, err, pid, go_pipe[2 ]; +union perf_event *event; +char bf; +int count =0 ; + +snprintf (path,PATH_MAX ,"%s/kernel/debug/tracing/events/syscalls/sys_enter_prctl/id" , +sysfs__mountpoint ()); + +if (filename__read_int (path, &id)) { +fprintf (stderr,"error: failed to get tracepoint id: %s\n" , path); +return -1 ; + } + + attr.config = id; + + err =pipe (go_pipe); +__T ("failed to create pipe" , err ==0 ); + +fflush (NULL ); + + pid =fork (); +if (!pid) { +int i; + +read (go_pipe[0 ], &bf,1 ); + +/* Generate 100 prctl calls. */ +for (i =0 ; i <100 ; i++) +prctl (0 ,0 ,0 ,0 ,0 ); + +exit (0 ); + } + + threads =perf_thread_map__new_dummy (); +__T ("failed to create threads" , threads); + + cpus =perf_cpu_map__dummy_new (); +__T ("failed to create cpus" , cpus); + +perf_thread_map__set_pid (threads,0 , pid); + + evlist =perf_evlist__new (); +__T ("failed to create evlist" , evlist); + + evsel =perf_evsel__new (&attr); +__T ("failed to create evsel1" , evsel); + +perf_evlist__add (evlist, evsel); + +perf_evlist__set_maps (evlist, cpus, threads); + + err =perf_evlist__open (evlist); +__T ("failed to open evlist" , err ==0 ); + + err =perf_evlist__mmap (evlist,4 ); +__T ("failed to mmap evlist" , err ==0 ); + +perf_evlist__enable (evlist); + +/* kick the child and wait for it to finish */ +write (go_pipe[1 ], &bf,1 ); +waitpid (pid,NULL ,0 ); + +/* +* There's no need to call perf_evlist__disable, +* monitored process is dead now. +*/ + +perf_evlist__for_each_mmap (evlist, map, false) { +if (perf_mmap__read_init (map) <0 ) +continue ; + +while ((event =perf_mmap__read_event (map)) !=NULL ) { + count++; +perf_mmap__consume (map); + } + +perf_mmap__read_done (map); + } + +/* calls perf_evlist__munmap/perf_evlist__close */ +perf_evlist__delete (evlist); + +perf_thread_map__put (threads); +perf_cpu_map__put (cpus); + +/* +* The generated prctl calls should match the +* number of events in the buffer. +*/ +__T ("failed count" , count ==100 ); + +return 0 ; +} +int main (int argc,char **argv) { __T_START;
@@ -190,6 +308,7 @@ int main(int argc, char **argv)test_stat_cpu ();test_stat_thread ();test_stat_thread_enable (); +test_mmap_thread (); __T_OK;return 0 ;
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 48/57 libperf: Add tests_mmap_cpus test
In reply to Arnaldo Carvalho de MeloFrom: Jiri Olsa <[email protected]>
Add mmaping tests that generates prctl call on every cpu validates it gets all the related events in ring buffer.
Committer testing:
# make -C tools/perf/lib tests
make: Entering directory '/home/acme/git/perf/tools/perf/lib'
LINK test-cpumap-a
LINK test-threadmap-a
LINK test-evlist-a
LINK test-evsel-a
LINK test-cpumap-so
LINK test-threadmap-so
LINK test-evlist-so
LINK test-evsel-so
running static:
- running test-cpumap.c...OK
- running test-threadmap.c...OK
- running test-evlist.c...OK
- running test-evsel.c...OK
running dynamic:
- running test-cpumap.c...OK
- running test-threadmap.c...OK
- running test-evlist.c...OK
- running test-evsel.c...OK
make: Leaving directory '/home/acme/git/perf/tools/perf/lib'
#[ Added _GNU_SOURCE define for sched.h to get sched_[gs]et_affinity
--- tools/perf/lib/tests/test-evlist.c | 98 ++++++++++++++++++++++++++++++ 1 file changed, 98 insertions(+)
tools/perf/lib/tests/test-evlist.c +98 -0
@@ -1,4 +1,6 @@// SPDX-License-Identifier: GPL-2.0 +#define _GNU_SOURCE// needed for sched.h to get sched_[gs]etaffinity and CPU_(ZERO,SET) +#include <sched.h> #include <stdio.h> #include <stdarg.h> #include <unistd.h>
@@ -299,6 +301,101 @@ static int test_mmap_thread(void)return 0 ; } +static int test_mmap_cpus (void ) +{ +struct perf_evlist *evlist; +struct perf_evsel *evsel; +struct perf_mmap *map; +struct perf_cpu_map *cpus; +struct perf_event_attr attr = { + .type =PERF_TYPE_TRACEPOINT , + .sample_period =1 , + .wakeup_watermark =1 , + .disabled =1 , + }; +cpu_set_t saved_mask; +char path[PATH_MAX ]; +int id, err, cpu, tmp; +union perf_event *event; +int count =0 ; + +snprintf (path,PATH_MAX ,"%s/kernel/debug/tracing/events/syscalls/sys_enter_prctl/id" , +sysfs__mountpoint ()); + +if (filename__read_int (path, &id)) { +fprintf (stderr,"error: failed to get tracepoint id: %s\n" , path); +return -1 ; + } + + attr.config = id; + + cpus =perf_cpu_map__new (NULL ); +__T ("failed to create cpus" , cpus); + + evlist =perf_evlist__new (); +__T ("failed to create evlist" , evlist); + + evsel =perf_evsel__new (&attr); +__T ("failed to create evsel1" , evsel); + +perf_evlist__add (evlist, evsel); + +perf_evlist__set_maps (evlist, cpus,NULL ); + + err =perf_evlist__open (evlist); +__T ("failed to open evlist" , err ==0 ); + + err =perf_evlist__mmap (evlist,4 ); +__T ("failed to mmap evlist" , err ==0 ); + +perf_evlist__enable (evlist); + + err =sched_getaffinity (0 ,sizeof (saved_mask), &saved_mask); +__T ("sched_getaffinity failed" , err ==0 ); + +perf_cpu_map__for_each_cpu (cpu, tmp, cpus) { +cpu_set_t mask; + +CPU_ZERO (&mask); +CPU_SET (cpu, &mask); + + err =sched_setaffinity (0 ,sizeof (mask), &mask); +__T ("sched_setaffinity failed" , err ==0 ); + +prctl (0 ,0 ,0 ,0 ,0 ); + } + + err =sched_setaffinity (0 ,sizeof (saved_mask), &saved_mask); +__T ("sched_setaffinity failed" , err ==0 ); + +perf_evlist__disable (evlist); + +perf_evlist__for_each_mmap (evlist, map, false) { +if (perf_mmap__read_init (map) <0 ) +continue ; + +while ((event =perf_mmap__read_event (map)) !=NULL ) { + count++; +perf_mmap__consume (map); + } + +perf_mmap__read_done (map); + } + +/* calls perf_evlist__munmap/perf_evlist__close */ +perf_evlist__delete (evlist); + +/* +* The generated prctl events should match the +* number of cpus or be bigger (we are system-wide). +*/ +__T ("failed count" , count >=perf_cpu_map__nr (cpus)); + +perf_cpu_map__put (cpus); + +return 0 ; +} +int main (int argc,char **argv) { __T_START;
@@ -309,6 +406,7 @@ int main(int argc, char **argv)test_stat_thread ();test_stat_thread_enable ();test_mmap_thread (); +test_mmap_cpus (); __T_OK;return 0 ;
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 49/57 libperf: Keep count of failed tests
In reply to Arnaldo Carvalho de MeloFrom: Jiri Olsa <[email protected]>
Keep the count of failed tests, so we get better output with failures, like:
# make tests
...
running static:
- running test-cpumap.c...OK
- running test-threadmap.c...OK
- running test-evlist.c...FAILED test-evlist.c:53 failed to create evsel2
FAILED test-evlist.c:163 failed to create evsel2
FAILED test-evlist.c:287 failed count
FAILED (3)
- running test-evsel.c...OK
running dynamic:
- running test-cpumap.c...OK
- running test-threadmap.c...OK
- running test-evlist.c...FAILED test-evlist.c:53 failed to create evsel2
FAILED test-evlist.c:163 failed to create evsel2
FAILED test-evlist.c:287 failed count
FAILED (3)
- running test-evsel.c...OK
...--- tools/perf/lib/include/internal/tests.h | 20 +++++++++++++++++--- tools/perf/lib/tests/test-cpumap.c | 2 +- tools/perf/lib/tests/test-evlist.c | 2 +- tools/perf/lib/tests/test-evsel.c | 2 +- tools/perf/lib/tests/test-threadmap.c | 2 +- 5 files changed, 21 insertions(+), 7 deletions(-)
tools/perf/lib/include/internal/tests.h +17 -3
@@ -4,14 +4,28 @@#include <stdio.h> -#define __T_START fprintf(stdout, "- running %s...", __FILE__) -#define __T_OK fprintf(stdout, "OK\n") -#define __T_FAIL fprintf(stdout, "FAIL\n") +int tests_failed; + +#define __T_START \ +do { \ + fprintf(stdout, "- running %s...", __FILE__); \ + fflush(NULL); \ + tests_failed = 0; \ +} while (0) + +#define __T_END \ +do { \ + if (tests_failed) \ + fprintf(stdout, " FAILED (%d)\n", tests_failed); \ + else \ + fprintf(stdout, "OK\n"); \ +} while (0)#define __T (text, cond) \ do { \ if (!(cond)) { \ fprintf(stderr, "FAILED %s:%d %s\n", __FILE__, __LINE__, text); \ + tests_failed++; \ return -1; \ } \ } while (0)
tools/perf/lib/tests/test-cpumap.c +1 -1
@@ -26,6 +26,6 @@ int main(int argc, char **argv)perf_cpu_map__put (cpus);perf_cpu_map__put (cpus); - __T_OK; + __T_END;return 0 ; }
tools/perf/lib/tests/test-evlist.c +1 -1
@@ -408,6 +408,6 @@ int main(int argc, char **argv)test_mmap_thread ();test_mmap_cpus (); - __T_OK; + __T_END;return 0 ; }
tools/perf/lib/tests/test-evsel.c +1 -1
@@ -130,6 +130,6 @@ int main(int argc, char **argv)test_stat_thread ();test_stat_thread_enable (); - __T_OK; + __T_END;return 0 ; }
tools/perf/lib/tests/test-threadmap.c +1 -1
@@ -26,6 +26,6 @@ int main(int argc, char **argv)perf_thread_map__put (threads);perf_thread_map__put (threads); - __T_OK; + __T_END;return 0 ; }
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 50/57 libperf: Do not export perf_evsel__init()/perf_evlist__init()
In reply to Arnaldo Carvalho de MeloFrom: Jiri Olsa <[email protected]>
There's no point in exporting perf_evsel__init()/perf_evlist__init(), it's called from perf_evsel__new()/perf_evlist__new() respectively.
It's used only from perf where perf_evsel()/perf_evlist() is embedded perf's evsel/evlist.
--- tools/perf/lib/include/internal/evlist.h | 1 + tools/perf/lib/include/internal/evsel.h | 1 + tools/perf/lib/include/perf/evlist.h | 1 - tools/perf/lib/include/perf/evsel.h | 2 -- tools/perf/lib/libperf.map | 2 -- 5 files changed, 2 insertions(+), 5 deletions(-)
tools/perf/lib/include/internal/evlist.h +1 -0
@@ -50,6 +50,7 @@ int perf_evlist__mmap_ops(struct perf_evlist *evlist,struct perf_evlist_mmap_ops *ops, struct perf_mmap_param *mp); +void perf_evlist__init (struct perf_evlist *evlist);void perf_evlist__exit (struct perf_evlist *evlist);/**
tools/perf/lib/include/internal/evsel.h +1 -0
@@ -50,6 +50,7 @@ struct perf_evsel {bool system_wide; }; +void perf_evsel__init (struct perf_evsel *evsel,struct perf_event_attr *attr);int perf_evsel__alloc_fd (struct perf_evsel *evsel,int ncpus,int nthreads);void perf_evsel__close_fd (struct perf_evsel *evsel);void perf_evsel__free_fd (struct perf_evsel *evsel);
tools/perf/lib/include/perf/evlist.h +0 -1
@@ -10,7 +10,6 @@ struct perf_evsel;struct perf_cpu_map ;struct perf_thread_map ; -LIBPERF_API void perf_evlist__init(struct perf_evlist *evlist); LIBPERF_API void perf_evlist__add(struct perf_evlist *evlist,struct perf_evsel *evsel); LIBPERF_API void perf_evlist__remove(struct perf_evlist *evlist,
tools/perf/lib/include/perf/evsel.h +0 -2
@@ -21,8 +21,6 @@ struct perf_counts_values { }; }; -LIBPERF_API void perf_evsel__init(struct perf_evsel *evsel, -struct perf_event_attr *attr); LIBPERF_API struct perf_evsel *perf_evsel__new(struct perf_event_attr *attr); LIBPERF_API void perf_evsel__delete(struct perf_evsel *evsel); LIBPERF_API int perf_evsel__open(struct perf_evsel *evsel, struct perf_cpu_map *cpus,
tools/perf/lib/libperf.map +0 -2
@@ -21,7 +21,6 @@ LIBPERF_0.0.1 { perf_evsel__delete; perf_evsel__enable; perf_evsel__disable; - perf_evsel__init; perf_evsel__open; perf_evsel__close; perf_evsel__read;
@@ -34,7 +33,6 @@ LIBPERF_0.0.1 { perf_evlist__close; perf_evlist__enable; perf_evlist__disable; - perf_evlist__init; perf_evlist__add; perf_evlist__remove; perf_evlist__next;
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 51/57 libperf: Add pr_err() macro
In reply to Arnaldo Carvalho de MeloFrom: Jiri Olsa <[email protected]>
And missing include for "perf/core.h" header, which provides LIBPERF_* debug levels and add missing pr_err() support.
--- tools/perf/lib/include/perf/core.h | 1 + tools/perf/lib/internal.h | 3 +++ 2 files changed, 4 insertions(+)
tools/perf/lib/include/perf/core.h +1 -0
@@ -9,6 +9,7 @@#endif enum libperf_print_level { +LIBPERF_ERR ,LIBPERF_WARN ,LIBPERF_INFO ,LIBPERF_DEBUG ,
tools/perf/lib/internal.h +3 -0
@@ -2,6 +2,8 @@#ifndef __LIBPERF_INTERNAL_H#define __LIBPERF_INTERNAL_H +#include <perf/core.h> +void libperf_print (enum libperf_print_level level,const char *format, ...) __attribute__((format (printf,2 ,3 )));
@@ -11,6 +13,7 @@ do { \ libperf_print(level, "libperf: " fmt, ##__VA_ARGS__); \ } while (0) +#define pr_err (fmt, ...) __pr(LIBPERF_ERR, fmt, ##__VA_ARGS__)#define pr_warning (fmt, ...) __pr(LIBPERF_WARN, fmt, ##__VA_ARGS__)#define pr_info (fmt, ...) __pr(LIBPERF_INFO, fmt, ##__VA_ARGS__)#define pr_debug (fmt, ...) __pr(LIBPERF_DEBUG, fmt, ##__VA_ARGS__)
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 52/57 libbeauty: Introduce syscall_arg__strtoul_strarrays()
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
To allow going from string to integer for 'struct strarrays'.
--- tools/perf/builtin-trace.c | 5 +++++ tools/perf/trace/beauty/beauty.h | 3 +++ 2 files changed, 8 insertions(+)
tools/perf/builtin-trace.c +5 -0
@@ -540,6 +540,11 @@ bool syscall_arg__strtoul_strarray(char *bf, size_t size, struct syscall_arg *arreturn strarray__strtoul (arg->parm, bf, size, ret); } +bool syscall_arg__strtoul_strarrays (char *bf,size_t size,struct syscall_arg *arg,u64 *ret) +{ +return strarrays__strtoul (arg->parm, bf, size, ret); +} +size_t syscall_arg__scnprintf_strarray_flags (char *bf,size_t size,struct syscall_arg *arg) {return strarray__scnprintf_flags (arg->parm, bf, size, arg->show_string_prefix, arg->val);
tools/perf/trace/beauty/beauty.h +3 -0
@@ -125,6 +125,9 @@ size_t syscall_arg__scnprintf_strarray_flags(char *bf, size_t size, struct syscabool syscall_arg__strtoul_strarray (char *bf,size_t size,struct syscall_arg *arg,u64 *ret);#define STUL_STRARRAY syscall_arg__strtoul_strarray +bool syscall_arg__strtoul_strarrays (char *bf,size_t size,struct syscall_arg *arg,u64 *ret); +#define STUL_STRARRAYS syscall_arg__strtoul_strarrays +size_t syscall_arg__scnprintf_x86_irq_vectors (char *bf,size_t size,struct syscall_arg *arg);#define SCA_X86_IRQ_VECTORS syscall_arg__scnprintf_x86_irq_vectors
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 53/57 perf trace: Use strtoul for the fcntl 'cmd' argument
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
Since its values are in two ranges of values we ended up codifying it using a 'struct strarrays', so now hook it up with STUL_STRARRAYS so that we can do:
# perf trace -e syscalls:*enter_fcntl --filter=cmd==SETLK||cmd==SETLKW
0.000 sssd_kcm/19021 syscalls:sys_enter_fcntl(fd: 13</var/lib/sss/secrets/secrets.ldb>, cmd: SETLK, arg: 0x7ffcf0a4dee0)
1.523 sssd_kcm/19021 syscalls:sys_enter_fcntl(fd: 13</var/lib/sss/secrets/secrets.ldb>, cmd: SETLK, arg: 0x7ffcf0a4de90)
1.629 sssd_kcm/19021 syscalls:sys_enter_fcntl(fd: 13</var/lib/sss/secrets/secrets.ldb>, cmd: SETLK, arg: 0x7ffcf0a4de90)
2.711 sssd_kcm/19021 syscalls:sys_enter_fcntl(fd: 13</var/lib/sss/secrets/secrets.ldb>, cmd: SETLK, arg: 0x7ffcf0a4de70)
^C#--- tools/perf/builtin-trace.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tools/perf/builtin-trace.c +2 -1
@@ -894,7 +894,8 @@ static struct syscall_fmt syscall_fmts[] = { { .name = "fchownat", .arg = { [0] = { .scnprintf = SCA_FDAT, /* fd */ }, }, }, { .name = "fcntl", - .arg = { [1] = { .scnprintf = SCA_FCNTL_CMD, /* cmd */ + .arg = { [1] = { .scnprintf = SCA_FCNTL_CMD, /* cmd */ + .strtoul = STUL_STRARRAYS, .parm = &strarrays__fcntl_cmds_arrays, .show_zero = true, }, [2] = { .scnprintf = SCA_FCNTL_ARG, /* arg */ }, }, },
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 54/57 libbeauty: Make the mmap_flags strarray visible outside of its beautifier
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
So that we can later use it with the strarray__strtoul_flags() routine that will be soon introduced.
--- tools/perf/trace/beauty/mmap.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
tools/perf/trace/beauty/mmap.c +2 -2
@@ -33,11 +33,11 @@ static size_t syscall_arg__scnprintf_mmap_prot(char *bf, size_t size,#define SCA_MMAP_PROT syscall_arg__scnprintf_mmap_prot -static size_t mmap__scnprintf_flags (unsigned long flags,char *bf,size_t size,bool show_prefix) -{#include "trace/beauty/generated/mmap_flags_array.c" static DEFINE_STRARRAY(mmap_flags, "MAP_"); +static size_t mmap__scnprintf_flags (unsigned long flags,char *bf,size_t size,bool show_prefix) +{return strarray__scnprintf_flags (&strarray__mmap_flags, bf, size, show_prefix, flags); }
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 55/57 libbeauty: Introduce strarray__strtoul_flags()
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
Counterpart of strarray__scnprintf_flags(), i.e. from a expression like:
# perf trace -e syscalls:sys_enter_mmap --filter="flags==PRIVATE|FIXED|DENYWRITE"
I.e. that "flags==PRIVATE|FIXED|DENYWRITE", turn that into
# perf trace -e syscalls:sys_enter_mmap --filter=0x812
--- tools/perf/builtin-trace.c | 45 +++++++++++++++++++++++++++++++- tools/perf/trace/beauty/beauty.h | 1 + 2 files changed, 45 insertions(+), 1 deletion(-)
tools/perf/builtin-trace.c +44 -1
@@ -586,6 +586,49 @@ bool strarray__strtoul(struct strarray *sa, char *bf, size_t size, u64 *ret)return false; } +bool strarray__strtoul_flags (struct strarray *sa,char *bf,size_t size,u64 *ret) +{ +u64 val =0 ; +char *tok = bf, *sep, *end; + + *ret =0 ; + +while (size !=0 ) { +int toklen = size; + + sep =memchr (tok,'|' , size); +if (sep !=NULL ) { + size -= sep - tok +1 ; + + end = sep -1 ; +while (end > tok &&isspace (*end)) + --end; + + toklen = end - tok +1 ; + } + +while (isspace (*tok)) + ++tok; + +if (isalpha (*tok) || *tok =='_' ) { +if (!strarray__strtoul (sa, tok, toklen, &val)) +return false; + }else { +bool is_hexa = tok[0 ] ==0 && (tok[1 ] ='x' || tok[1 ] =='X' ); + + val =strtoul (tok,NULL , is_hexa ?16 :0 ); + } + + *ret |= (1 << (val -1 )); + +if (sep ==NULL ) +break ; + tok = sep +1 ; + } + +return true; +} +bool strarrays__strtoul (struct strarrays *sas,char *bf,size_t size,u64 *ret) {int i;
@@ -3676,7 +3719,7 @@ static int trace__expand_filter(struct trace *trace __maybe_unused, struct evsel } right_end = right +1 ; -while (isalnum (*right_end) || *right_end =='_' ) +while (isalnum (*right_end) || *right_end =='_' || *right_end =='|' ) ++right_end;if (isalpha (*right)) {
tools/perf/trace/beauty/beauty.h +1 -0
@@ -32,6 +32,7 @@ size_t strarray__scnprintf_suffix(struct strarray *sa, char *bf, size_t size, cosize_t strarray__scnprintf_flags (struct strarray *sa,char *bf,size_t size,bool show_prefix,unsigned long flags);bool strarray__strtoul (struct strarray *sa,char *bf,size_t size,u64 *ret); +bool strarray__strtoul_flags (struct strarray *sa,char *bf,size_t size,u64 *ret);struct trace ;struct thread ;
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 56/57 perf trace: Wire up strarray__strtoul_flags()
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
Now anything that uses STRARRAY_FLAGS, like the 'fsmount' syscall will support mapping or-ed strings back to a value that can be used in a filter.
In some cases, where STRARRAY_FLAGS isn't used but instead the scnprintf is a special one because of specific needs, like for mmap, then one has to set the ->pars to the strarray. See the next cset.
--- tools/perf/builtin-trace.c | 6 ++++++ tools/perf/trace/beauty/beauty.h | 3 +++ 2 files changed, 9 insertions(+)
tools/perf/builtin-trace.c +6 -0
@@ -540,6 +540,11 @@ bool syscall_arg__strtoul_strarray(char *bf, size_t size, struct syscall_arg *arreturn strarray__strtoul (arg->parm, bf, size, ret); } +bool syscall_arg__strtoul_strarray_flags (char *bf,size_t size,struct syscall_arg *arg,u64 *ret) +{ +return strarray__strtoul_flags (arg->parm, bf, size, ret); +} +bool syscall_arg__strtoul_strarrays (char *bf,size_t size,struct syscall_arg *arg,u64 *ret) {return strarrays__strtoul (arg->parm, bf, size, ret);
@@ -882,6 +887,7 @@ static size_t syscall_arg__scnprintf_getrandom_flags(char *bf, size_t size,#define STRARRAY_FLAGS (name, array) \ { .scnprintf = SCA_STRARRAY_FLAGS, \ + .strtoul = STUL_STRARRAY_FLAGS, \ .parm = &strarray__##array, }#include "trace/beauty/arch_errno_names.c"
tools/perf/trace/beauty/beauty.h +3 -0
@@ -126,6 +126,9 @@ size_t syscall_arg__scnprintf_strarray_flags(char *bf, size_t size, struct syscabool syscall_arg__strtoul_strarray (char *bf,size_t size,struct syscall_arg *arg,u64 *ret);#define STUL_STRARRAY syscall_arg__strtoul_strarray +bool syscall_arg__strtoul_strarray_flags (char *bf,size_t size,struct syscall_arg *arg,u64 *ret); +#define STUL_STRARRAY_FLAGS syscall_arg__strtoul_strarray_flags +bool syscall_arg__strtoul_strarrays (char *bf,size_t size,struct syscall_arg *arg,u64 *ret);#define STUL_STRARRAYS syscall_arg__strtoul_strarrays
Signature · 1 line
2.21.0
Arnaldo Carvalho de Melo
maintainerPATCH 57/57 perf trace: Use STUL_STRARRAY_FLAGS with mmap
In reply to Arnaldo Carvalho de MeloFrom: Arnaldo Carvalho de Melo <[email protected]>
The 'mmap' syscall has special needs so it doesn't use SCA_STRARRAY_FLAGS, see its implementation in syscall_arg__scnprintf_mmap_flags(), related to special handling of MAP_ANONYMOUS, so set ->parm to the strarray__mmap_flags and hook up with strarray__strtoul_flags manually, now we can filter by those or-ed string expressions:
# perf trace -e syscalls:sys_enter_mmap sleep 1
0.000 syscalls:sys_enter_mmap(addr: NULL, len: 134346, prot: READ, flags: PRIVATE, fd: 3, off: 0)
0.026 syscalls:sys_enter_mmap(addr: NULL, len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS)
0.036 syscalls:sys_enter_mmap(addr: NULL, len: 1857472, prot: READ, flags: PRIVATE|DENYWRITE, fd: 3, off: 0)
0.046 syscalls:sys_enter_mmap(addr: 0x7fae003d9000, len: 1363968, prot: READ|EXEC, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x22000)
0.052 syscalls:sys_enter_mmap(addr: 0x7fae00526000, len: 311296, prot: READ, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x16f000)
0.055 syscalls:sys_enter_mmap(addr: 0x7fae00573000, len: 24576, prot: READ|WRITE, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x1bb000)
0.062 syscalls:sys_enter_mmap(addr: 0x7fae00579000, len: 14272, prot: READ|WRITE, flags: PRIVATE|FIXED|ANONYMOUS)
0.253 syscalls:sys_enter_mmap(addr: NULL, len: 217750512, prot: READ, flags: PRIVATE, fd: 3, off: 0)
# # perf trace -e syscalls:sys_enter_mmap --filter="flags==PRIVATE|FIXED|DENYWRITE" sleep 1
0.000 syscalls:sys_enter_mmap(addr: 0x7f6ab3dcb000, len: 1363968, prot: READ|EXEC, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x22000)
0.010 syscalls:sys_enter_mmap(addr: 0x7f6ab3f18000, len: 311296, prot: READ, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x16f000)
0.014 syscalls:sys_enter_mmap(addr: 0x7f6ab3f65000, len: 24576, prot: READ|WRITE, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x1bb000)
# perf trace -e syscalls:sys_enter_mmap --filter="flags==PRIVATE|ANONYMOUS" sleep 1
0.000 syscalls:sys_enter_mmap(addr: NULL, len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS)
## perf trace -v -e syscalls:sys_enter_mmap --filter="flags==PRIVATE|ANONYMOUS" sleep 1 |& grep "New filter" New filter for syscalls:sys_enter_mmap: flags==0x22 #
--- tools/perf/builtin-trace.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tools/perf/builtin-trace.c +3 -1
@@ -1015,7 +1015,9 @@ static struct syscall_fmt syscall_fmts[] = { .alias = "old_mmap", #endif .arg = { [2] = { .scnprintf = SCA_MMAP_PROT, /* prot */ }, - [3] = { .scnprintf = SCA_MMAP_FLAGS, /* flags */ }, + [3] = { .scnprintf = SCA_MMAP_FLAGS, /* flags */ + .strtoul = STUL_STRARRAY_FLAGS, + .parm = &strarray__mmap_flags, }, [5] = { .scnprintf = SCA_HEX, /* offset */ }, }, }, { .name = "mount", .arg = { [0] = { .scnprintf = SCA_FILENAME, /* dev_name */ },
Signature · 1 line
2.21.0
Quoted 352 lines · Arnaldo Carvalho de Melo
Hi Ingo/Thomas,
Please consider pulling,
Best regards,
- Arnaldo
Test results at the end of this message, as usual.
The following changes since commit
39b656ee9f2ce41eb969c86525f9a2a63fefac5b:Merge tag 'perf-core-for-mingo-5.5-20191011' of
git://git.kernel.org/pub/scm/linux/kernel/git/acme/linuxinto perf/core (2019-10-15 07:19:55 +0200)are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.gittags/perf-core-for-mingo-5.5-20191021for you to fetch changes up to
27198a893ba074407e7a87e346252b3e6fab454f:perf trace: Use
STUL_STRARRAY_FLAGSwith mmap (2019-10-19 15:35:02 -0300)----------------------------------------------------------------
perf/core improvements and fixes:perf trace:
- Add syscall failure stats to -s/--summary and -S/--with-summary, also works in combination with specifying just a set of syscalls, see below first with -s/--summary, then with -S/--with-summary just for the syscalls we saw failing with -s:
# perf trace -s sleep 1Summary of events: sleep (16218), 80 events, 93.0% syscall calls errors total min avg max stddev (msec) (msec) (msec) (msec) (%) ----------- ----- ------ -------- -------- -------- -------- ------ nanosleep 1 0 1000.091 1000.091 1000.091 1000.091 0.00% mmap 8 0 0.045 0.005 0.006 0.008 7.09% mprotect 4 0 0.028 0.005 0.007 0.009 11.38% openat 3 0 0.021 0.005 0.007 0.009 14.07% munmap 1 0 0.017 0.017 0.017 0.017 0.00% brk 4 0 0.010 0.001 0.002 0.004 23.15% read 4 0 0.009 0.002 0.002 0.003 8.13% close 5 0 0.008 0.001 0.002 0.002 10.83% fstat 3 0 0.006 0.002 0.002 0.002 6.97% access 1 1 0.006 0.006 0.006 0.006 0.00% lseek 3 0 0.005 0.001 0.002 0.002 7.37% arch_prctl 2 1 0.004 0.001 0.002 0.002 17.64% execve 1 0 0.000 0.000 0.000 0.000 0.00% # perf trace -e access,arch_prctl -S sleep 1 0.000 ( 0.006 ms): sleep/19503 arch_prctl(option: 0x3001, arg2: 0x7fff165996b0) = -1 EINVAL (Invalid argument) 0.024 ( 0.006 ms): sleep/19503 access(filename: 0x2177e510, mode: R) = -1 ENOENT (No such file or directory) 0.136 ( 0.002 ms): sleep/19503 arch_prctl(option: SET_FS, arg2: 0x7f9421737580) = 0 Summary of events: sleep (19503), 6 events, 50.0% syscall calls errors total min avg max stddev (msec) (msec) (msec) (msec) (%) ---------- ----- ------ ------ ------ ------ ------ ------ arch_prctl 2 1 0.008 0.002 0.004 0.006 57.22% access 1 1 0.006 0.006 0.006 0.006 0.00% #- Introduce --errno-summary, to drill down a bit more in the errno stats:
# perf trace --errno-summary -e access,arch_prctl -S sleep 1 0.000 ( 0.006 ms): sleep/5587 arch_prctl(option: 0x3001, arg2: 0x7ffd6ba6aa00) = -1 EINVAL (Invalid argument) 0.028 ( 0.007 ms): sleep/5587 access(filename: 0xb83d9510, mode: R) = -1 ENOENT (No such file or directory) 0.172 ( 0.003 ms): sleep/5587 arch_prctl(option: SET_FS, arg2: 0x7f45b8392580) = 0Summary of events: sleep (5587), 6 events, 50.0% syscall calls errors total min avg max stddev (msec) (msec) (msec) (msec) (%) ---------- ----- ------ ------ ------ ------ ------ ------ arch_prctl 2 1 0.009 0.003 0.005 0.006 38.90% EINVAL: 1 access 1 1 0.007 0.007 0.007 0.007 0.00% ENOENT: 1 #- Filter own pid to avoid a feedback look in 'perf trace record -a'
- Add the glue for the auto generated x86 IRQ vector array.
- Show error message when not finding a field used in a filter expression
# perf trace --max-events=4 -e syscalls:sys_enter_write --filter="cnt>32767" Failed to set filter "(cnt>32767) && (common_pid != 19938 && common_pid != 8922)" on event syscalls:sys_enter_write with 22 (Invalid argument) # # perf trace --max-events=4 -e syscalls:sys_enter_write --filter="count>32767" 0.000 python3.5/17535 syscalls:sys_enter_write(fd: 3, buf: 0x564b0dc53600, count: 172086) 12.641 python3.5.post/17535 syscalls:sys_enter_write(fd: 3, buf: 0x564b0db63660, count: 75994) 27.738 python3.5.post/17535 syscalls:sys_enter_write(fd: 3, buf: 0x564b0db4b1e0, count: 41635) 136.070 python3.5.post/17535 syscalls:sys_enter_write(fd: 3, buf: 0x564b0dbab510, count: 62232) #- Add a generator for x86's IRQ vectors -> strings
- Introduce
stroul()(string -> number) methods for the strarray andstrarrays classes, also strtoul_flags, allowing to go from both strings and or-ed strings to numbers, allowing things like: # perf trace -e syscalls:sys_enter_mmap --filter="flags==DENYWRITE|PRIVATE|FIXED" sleep 1 0.000 sleep/22588 syscalls:sys_enter_mmap(addr: 0x7f42d2aa5000, len: 1363968, prot: READ|EXEC, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x22000) 0.011 sleep/22588 syscalls:sys_enter_mmap(addr: 0x7f42d2bf2000, len: 311296, prot: READ, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x16f000) 0.015 sleep/22588 syscalls:sys_enter_mmap(addr: 0x7f42d2c3f000, len: 24576, prot: READ|WRITE, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x1bb000) #Allowing to narrow down from the complete set of mmap calls for that workload:
# perf trace -e syscalls:sys_enter_mmap sleep 1 0.000 sleep/22695 syscalls:sys_enter_mmap(len: 134773, prot: READ, flags: PRIVATE, fd: 3) 0.041 sleep/22695 syscalls:sys_enter_mmap(len: 8192, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS) 0.053 sleep/22695 syscalls:sys_enter_mmap(len: 1857472, prot: READ, flags: PRIVATE|DENYWRITE, fd: 3) 0.069 sleep/22695 syscalls:sys_enter_mmap(addr: 0x7fd23ffb6000, len: 1363968, prot: READ|EXEC, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x22000) 0.077 sleep/22695 syscalls:sys_enter_mmap(addr: 0x7fd240103000, len: 311296, prot: READ, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x16f000) 0.083 sleep/22695 syscalls:sys_enter_mmap(addr: 0x7fd240150000, len: 24576, prot: READ|WRITE, flags: PRIVATE|FIXED|DENYWRITE, fd: 3, off: 0x1bb000) 0.095 sleep/22695 syscalls:sys_enter_mmap(addr: 0x7fd240156000, len: 14272, prot: READ|WRITE, flags: PRIVATE|FIXED|ANONYMOUS) 0.339 sleep/22695 syscalls:sys_enter_mmap(len: 217750512, prot: READ, flags: PRIVATE, fd: 3) #Works with all targets, so, for system wide, looking at who calls mmap with flags set to just "PRIVATE":
# perf trace --max-events=5 -e syscalls:sys_enter_mmap --filter="flags==PRIVATE" 0.000 pool/2242 syscalls:sys_enter_mmap(len: 756, prot: READ, flags: PRIVATE, fd: 14) 0.050 pool/2242 syscalls:sys_enter_mmap(len: 756, prot: READ, flags: PRIVATE, fd: 14) 0.062 pool/2242 syscalls:sys_enter_mmap(len: 756, prot: READ, flags: PRIVATE, fd: 14) 0.145 goa-identity-s/2240 syscalls:sys_enter_mmap(len: 756, prot: READ, flags: PRIVATE, fd: 18) 0.183 goa-identity-s/2240 syscalls:sys_enter_mmap(len: 756, prot: READ, flags: PRIVATE, fd: 18) ## perf trace --max-events=2 -e syscalls:sys_enter_lseek --filter="whence==SET && offset != 0" 0.000 Cache2 I/O/12047 syscalls:sys_enter_lseek(fd: 277, offset: 43, whence: SET) 1142.070 mozStorage #5/12302 syscalls:sys_enter_lseek(fd: 44</home/acme/.mozilla/firefox/ina67tev.default/cookies.sqlite-wal>, offset: 393536, whence: SET) #perf annotate:
- Fix objdump --no-show-raw-insn flag to work with goth gcc and clang.
- Streamline objdump execution, preserving the right error codes for better reporting to user.
perf report:
- Add warning when libunwind not compiled in.
perf stat:
Jin Yao:
- Support --all-kernel/--all-user, to match options available in 'perf record', asking that all the events specified work just with kernel or user events.
perf list:
Jin Yao:
- Hide deprecated events by default, allow showing them with --deprecated.
libbperf:
Jiri Olsa:
- Allow to build with -ltcmalloc.
- Finish mmap interface, getting more stuff from
tools/perfwhile adding abstractions to avoid pulling too much stuff, to get libperf to grow as tools needs things like auxtrace, etc.perf scripting engines:
Steven Rostedt (VMware):
- Iterate on tep event arrays directly, fixing script generation with '-g python' when having multiple tracepoints in a perf.data file.
core:
- Allow to build with -ltcmalloc.
perf test:
Leo Yan:
- Report failure for mmap events.
- Avoid infinite loop for task exit case.
- Remove needless headers for
bp_accounttest.- Add dedicated checking helper
is_supported().- Disable
bp_signaltesting for arm64.Vendor events:
arm64:
John Garry:
- Fix Hisi hip08 DDRC PMU eventname.
- Add some missing events for Hisi hip08 DDRC, L3C and HHA PMUs.
Signed-off-by Arnaldo Carvalho de Melo <[email protected]>----------------------------------------------------------------
Andi Kleen (2):
perf script: Fix --reltime with --time
perf evlist: Fix fix for freed id arraysArnaldo Carvalho de Melo (25):
perf trace: Add syscall failure stats to -s/--summary and -S/--with-summary perf trace: Introduce --errno-summary
perf string: Exportasprintf__tp_filter_pids()
perf trace: Filter own pid to avoid a feedback look in 'perf trace record -a' perf trace: Support tracepoint dynamic char arrays
tools arch x86: Grab a copy of the file containing the IRQ vector defines
libbeauty: Add a generator for x86's IRQ vectors -> strings
libbeauty: Hook up the x86irq_vectorstable generator
libbeauty: Add astrarray__scnprintf_suffix()method
perf trace beauty: Add the glue for the autogenerated x86 IRQ vector array perf trace: Hook the 'vec' tracepoint argument with the x86 IRQ vectors scnprintf/strtoul perf trace: Show error message when not finding a field used in a filter expression perf trace: Introduce accessors to trace specificevsel->priv
perf trace: Hideevsel->accessfurther, simplify code
perf trace: Introduce 'structevsel__trace' forevsel->privneeds
perf trace: Initializeevsel_trace->fmtfor syscalls:sys_enter_* tracepoints
libbeauty: Introducesyscall_arg__strtoul_strarray()
perf trace: Honour --max-events in processing syscalls:sys_enter_*
perf trace: Pass asyscall_argtosyscall_arg_fmt->strtoul()
libbeauty: Introducesyscall_arg__strtoul_strarrays()
perf trace: Use strtoul for the fcntl 'cmd' argument
libbeauty: Make themmap_flagsstrarray visible outside of its beautifier
libbeauty: Introducestrarray__strtoul_flags()
perf trace: Wire upstrarray__strtoul_flags()
perf trace: UseSTUL_STRARRAY_FLAGSwith mmapIan Rogers (5):
perf annotate: Avoid reallocation in objdump parsing perf annotate: Use libsubcmd'srun-command.hto fork objdump perf annotate: Don't pipe objdump output through 'grep' command perf annotate: Don't pipe objdump output through 'expand' command perf annotate: Fix objdump --no-show-raw-insn flagJin Yao (3):
perf report: Add warning when libunwind not compiled in perf stat: Support --all-kernel/--all-user perf list: Hide deprecated events by defaultJiri Olsa (10):
perf tools: Allow to build with -ltcmalloc
libperf: Introduceperf_evlist__for_each_mmap()
libperf: Move mmap allocation toperf_evlist__mmap_ops::get
libperf: Move mask setup toperf_evlist__mmap_ops()
libperf: Link static tests with libapi.a
libperf: Addtests_mmap_threadtest
libperf: Addtests_mmap_cpustest
libperf: Keep count of failed tests
libperf: Do not export perf_evsel__init()/perf_evlist__init()
libperf: Addpr_err()macroJohn Garry (4):
perf vendor events arm64: Fix Hisi hip08 DDRC PMU eventname perf vendor events arm64: Add some missing events for Hisi hip08 DDRC PMU perf vendor events arm64: Add some missing events for Hisi hip08 L3C PMU perf vendor events arm64: Add some missing events for Hisi hip08 HHA PMULeo Yan (5):
perf test: Report failure for mmap events
perf test: Avoid infinite loop for task exit case
perf tests: Remove needless headers forbp_account
perf testsbp_account: Add dedicated checking helperis_supported()perf tests: Disablebp_signaltesting for arm64Steven Rostedt (VMware) (2):
perf scripting engines: Iterate on tep event arrays directly perf tools: Remove unusedtrace_find_next_event()Thomas Richter (1):
perf jvmti: Link againsttools/lib/ctype.hto have weakstrlcpy()tools/arch/x86/include/asm/irq_vectors.h | 146 +++++++ tools/perf/Documentation/perf-list.txt | 3 + tools/perf/Documentation/perf-stat.txt | 6 + tools/perf/Documentation/perf-trace.txt | 4 + tools/perf/Makefile.config | 5 + tools/perf/Makefile.perf | 10 + tools/perf/builtin-list.c | 14 +- tools/perf/builtin-report.c | 7 + tools/perf/builtin-script.c | 5 +- tools/perf/builtin-stat.c | 6 + tools/perf/builtin-trace.c | 420 ++++++++++++++++----- tools/perf/check-headers.sh | 1 + tools/perf/jvmti/Build | 6 +- tools/perf/lib/Makefile | 1 + tools/perf/lib/evlist.c | 71 +++- tools/perf/lib/include/internal/evlist.h | 3 + tools/perf/lib/include/internal/evsel.h | 1 + tools/perf/lib/include/internal/mmap.h | 5 +- tools/perf/lib/include/internal/tests.h | 20 +- tools/perf/lib/include/perf/core.h | 1 + tools/perf/lib/include/perf/evlist.h | 10 +- tools/perf/lib/include/perf/evsel.h | 2 - tools/perf/lib/internal.h | 3 + tools/perf/lib/libperf.map | 3 +- tools/perf/lib/mmap.c | 6 +- tools/perf/lib/tests/Makefile | 6 +- tools/perf/lib/tests/test-cpumap.c | 2 +- tools/perf/lib/tests/test-evlist.c | 219 ++++++++++- tools/perf/lib/tests/test-evsel.c | 2 +- tools/perf/lib/tests/test-threadmap.c | 2 +- .../arch/arm64/hisilicon/hip08/uncore-ddrc.json | 16 +- .../arch/arm64/hisilicon/hip08/uncore-hha.json | 23 +- .../arch/arm64/hisilicon/hip08/uncore-l3c.json | 56 +++ tools/perf/pmu-events/jevents.c | 26 +- tools/perf/pmu-events/jevents.h | 3 +- tools/perf/pmu-events/pmu-events.h | 1 + tools/perf/tests/bp_account.c | 20 +- tools/perf/tests/bp_signal.c | 15 +- tools/perf/tests/builtin-test.c | 2 +- tools/perf/tests/task-exit.c | 9 + tools/perf/tests/tests.h | 1 + tools/perf/trace/beauty/beauty.h | 19 + tools/perf/trace/beauty/mmap.c | 4 +- tools/perf/trace/beauty/tracepoints/Build | 1 + .../trace/beauty/tracepoints/x86_irq_vectors.c | 29 ++ .../trace/beauty/tracepoints/x86_irq_vectors.sh | 27 ++ tools/perf/util/annotate.c | 196 ++++++---- tools/perf/util/evlist.c | 34 +- tools/perf/util/parse-events.c | 4 +- tools/perf/util/parse-events.h | 2 +- tools/perf/util/pmu.c | 17 +- tools/perf/util/pmu.h | 4 +- .../perf/util/scripting-engines/trace-event-perl.c | 8 +- .../util/scripting-engines/trace-event-python.c | 9 +- tools/perf/util/stat.c | 10 + tools/perf/util/stat.h | 2 + tools/perf/util/string2.h | 3 +
Quoted 8 lines
tools/perf/util/time-utils.c | 27 +- tools/perf/util/time-utils.h | 5 + tools/perf/util/trace-event-parse.c | 31 -- tools/perf/util/trace-event.h | 2 - 61 files changed, 1307 insertions(+), 289 deletions(-) create mode 100644 tools/arch/x86/include/asm/irq_vectors.h create mode 100644 tools/perf/trace/beauty/tracepoints/x86_irq_vectors.c create mode 100755 tools/perf/trace/beauty/tracepoints/x86_irq_vectors.sh
Pulled, thanks a lot Arnaldo!
Ingo