And move the existing one to fsmount_attr.sh to be more precise.
Now the fsmount_flags[] is generated from the mount.h like below.
The ilog2() + 1 is an existing pattern to handle bit flags.
$ cat tools/perf/trace/beauty/generated/fsmount_arrays.c
static const char *fsmount_flags[] = {
[ilog2(0x00000001) + 1] = "CLOEXEC",
[ilog2(0x00000002) + 1] = "NAMESPACE",
};
It was found by Sashiko during the review.
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Don't print header differences during the perf build as it's noisy.
Mostly people won't care and find it annoying.
As it's to improve perf trace beautifier to catch up new changes mostly
in UAPIs, we can make it a separate build target and call it
occasionally. Make it and build-test related targets phony.
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
As it is not really used when compiling anything, just being parsed to
collect number->string tables for 'perf trace'.
$ git grep fadvise.h tools/
tools/perf/Makefile.perf:$(fadvise_advice_array): $(beauty_uapi_linux_dir)/fadvise.h $(fadvise_advice_tbl)
tools/perf/check-headers.sh: "include/uapi/linux/fadvise.h"
tools/perf/trace/beauty/fadvise.sh:grep -E $regex ${header_dir}/fadvise.h | \
tools/perf/trace/beauty/fadvise.sh:# tools/include/uapi/linux/fadvise.h for details.
$
Link: https://lore.kernel.org/r/CAP-5=fVBNQVF8k3JUQjH1nkP69ZVp8BqP+uwygcx=xO0zC4xrg@mail.gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
As it is used only to parse ioctl numbers, not to build perf and so far
no other tools/ living tool uses it, so to clean up tools/include/ to be
used just for building tools, to have access to things available in the
kernel and not yet in the system headers, move it to the directory where
just the tools/perf/trace/beauty/ scripts can use to generate tables
used by perf.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Currently no target is specified to compile rust code when needed, which
breaks cross compilation. E.g. for arm64:
LD /tmp/build/tests/workloads/perf-test-in.o
aarch64-linux-gnu-ld: /tmp/build/tests/workloads/code_with_type.a(code_with_type.code_with_type.d12f4324cb53c560-cgu.0.rcgu.o): Relocations in generic ELF (EM: 62)
aarch64-linux-gnu-ld: /tmp/build/tests/workloads/code_with_type.a(code_with_type.code_with_type.d12f4324cb53c560-cgu.0.rcgu.o): Relocations in generic ELF (EM: 62)
[...repeated...]
aarch64-linux-gnu-ld: /tmp/build/tests/workloads/code_with_type.a(code_with_type.code_with_type.d12f4324cb53c560-cgu.0.rcgu.o): Relocations in generic ELF (EM: 62)
aarch64-linux-gnu-ld: /tmp/build/tests/workloads/code_with_type.a(code_with_type.code_with_type.d12f4324cb53c560-cgu.0.rcgu.o): Relocations in generic ELF (EM: 62)
aarch64-linux-gnu-ld: /tmp/build/tests/workloads/code_with_type.a: error adding symbols: file in wrong format
make[5]: *** [/perf/tools/build/Makefile.build:162: /tmp/build/tests/workloads/perf-test-in.o] Error 1
make[4]: *** [/perf/tools/build/Makefile.build:156: workloads] Error 2
make[3]: *** [/perf/tools/build/Makefile.build:156: tests] Error 2
make[2]: *** [Makefile.perf:785: /tmp/build/perf-test-in.o] Error 2
make[2]: *** Waiting for unfinished jobs....
make[1]: *** [Makefile.perf:289: sub-make] Error 2
make: *** [Makefile:76: all] Error 2
Detect required target and pass it via rust_flags to the compiler.
Note that CROSS_COMPILE might be different from what rust compiler
expects, since it may omit the target vendor value, e.g.
"aarch64-linux-gnu" instead of "aarch64-unknown-linux-gnu".
Thus explicitly map supported CROSS_COMPILE values to corresponding Rust
versions, as suggested by Miguel Ojeda.
Tested using arm64 cross-compilation example from [1].
Fixes: 2e05bb52a1 ("perf test workload: Add code_with_type test workload")
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Dmitrii Dolgov <9erthalion6@gmail.com>
Cc: Levi Zim <i@kxxt.dev>
Cc: Miguel Ojeda <ojeda@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Nathan Chancellor <nathan@kernel.org>
Cc: Nicolas Schier <nsc@kernel.org>
Link: https://perfwiki.github.io/main/arm64-cross-compilation-dockerfile/ [1]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The purpose of the workload is to gather samples of rust runtime. To
achieve that it has a dummy rust library linked with it.
Per recommendations for such scenarios [1], the rust library is
statically linked.
An example:
$ perf record perf test -w code_with_type
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.160 MB perf.data (4074 samples) ]
$ perf report --stdio --dso perf -s srcfile,srcline
45.16% ub_checks.rs ub_checks.rs:72
6.72% code_with_type.rs code_with_type.rs:15
6.64% range.rs range.rs:767
4.26% code_with_type.rs code_with_type.rs:21
4.23% range.rs range.rs:0
3.99% code_with_type.rs code_with_type.rs:16
[...]
[1]: https://doc.rust-lang.org/reference/linkage.html#mixed-rust-and-foreign-codebases
Signed-off-by: Dmitrii Dolgov <9erthalion6@gmail.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Pull perf tools updates from Namhyung Kim:
"Perf event/metric description:
Unify all event and metric descriptions in JSON format. Now event
parsing and handling is greatly simplified by that.
From users point of view, perf list will provide richer information
about hardware events like the following.
$ perf list hw
List of pre-defined events (to be used in -e or -M):
legacy hardware:
branch-instructions
[Retired branch instructions [This event is an alias of branches]. Unit: cpu]
branch-misses
[Mispredicted branch instructions. Unit: cpu]
branches
[Retired branch instructions [This event is an alias of branch-instructions]. Unit: cpu]
bus-cycles
[Bus cycles,which can be different from total cycles. Unit: cpu]
cache-misses
[Cache misses. Usually this indicates Last Level Cache misses; this is intended to be used in conjunction with the
PERF_COUNT_HW_CACHE_REFERENCES event to calculate cache miss rates. Unit: cpu]
cache-references
[Cache accesses. Usually this indicates Last Level Cache accesses but this may vary depending on your CPU. This may include
prefetches and coherency messages; again this depends on the design of your CPU. Unit: cpu]
cpu-cycles
[Total cycles. Be wary of what happens during CPU frequency scaling [This event is an alias of cycles]. Unit: cpu]
cycles
[Total cycles. Be wary of what happens during CPU frequency scaling [This event is an alias of cpu-cycles]. Unit: cpu]
instructions
[Retired instructions. Be careful,these can be affected by various issues,most notably hardware interrupt counts. Unit: cpu]
ref-cycles
[Total cycles; not affected by CPU frequency scaling. Unit: cpu]
But most notable changes would be in the perf stat. On the right side,
the default metrics are better named and aligned. :)
$ perf stat -- perf test -w noploop
Performance counter stats for 'perf test -w noploop':
11 context-switches # 10.8 cs/sec cs_per_second
0 cpu-migrations # 0.0 migrations/sec migrations_per_second
3,612 page-faults # 3532.5 faults/sec page_faults_per_second
1,022.51 msec task-clock # 1.0 CPUs CPUs_utilized
110,466 branch-misses # 0.0 % branch_miss_rate (88.66%)
6,934,452,104 branches # 6781.8 M/sec branch_frequency (88.66%)
4,657,032,590 cpu-cycles # 4.6 GHz cycles_frequency (88.65%)
27,755,874,218 instructions # 6.0 instructions insn_per_cycle (89.03%)
TopdownL1 # 0.3 % tma_backend_bound
# 9.3 % tma_bad_speculation (89.05%)
# 9.7 % tma_frontend_bound (77.86%)
# 80.7 % tma_retiring (88.81%)
1.025318171 seconds time elapsed
1.013248000 seconds user
0.012014000 seconds sys
Deferred unwinding support:
With the kernel support (commit c69993ecdd: "perf: Support deferred
user unwind"), perf can use deferred callchains for userspace stack
trace with frame pointers like below:
$ perf record --call-graph fp,defer ...
This will be transparent to users when it comes to other commands like
perf report and perf script. They will merge the deferred callchains
to the previous samples as if they were collected together.
ARM SPE updates
- Extensive enhancements to support various kinds of memory
operations including GCS, MTE allocation tags, memcpy/memset,
register access, and SIMD operations.
- Add inverted data source filter (inv_data_src_filter) support to
exclude certain data sources.
- Improve documentation.
Vendor event updates:
- Intel: Updated event files for Sierra Forest, Panther Lake, Meteor
Lake, Lunar Lake, Granite Rapids, and others.
- Arm64: Added metrics for i.MX94 DDR PMU and Cortex-A720AE
definitions.
- RISC-V: Added JSON support for T-HEAD C920V2.
Misc:
- Improve pointer tracking in data type profiling. It'd give better
output when the variable is using container_of() to convert type.
- Annotation support for perf c2c report in TUI. Press 'a' key to
enter annotation view from cacheline browser window. This will show
which instruction is causing the cacheline contention.
- Lots of fixes and test coverage improvements!"
* tag 'perf-tools-for-v6.19-2025-12-06' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools: (214 commits)
libperf: Use 'extern' in LIBPERF_API visibility macro
perf stat: Improve handling of termination by signal
perf tests stat: Add test for error for an offline CPU
perf stat: When no events, don't report an error if there is none
perf tests stat: Add "--null" coverage
perf cpumap: Add "any" CPU handling to cpu_map__snprint_mask
libperf cpumap: Fix perf_cpu_map__max for an empty/NULL map
perf stat: Allow no events to open if this is a "--null" run
perf test kvm: Add some basic perf kvm test coverage
perf tests evlist: Add basic evlist test
perf tests script dlfilter: Add a dlfilter test
perf tests kallsyms: Add basic kallsyms test
perf tests timechart: Add a perf timechart test
perf tests top: Add basic perf top coverage test
perf tests buildid: Add purge and remove testing
perf tests c2c: Add a basic c2c
perf c2c: Clean up some defensive gets and make asan clean
perf jitdump: Fix missed dso__put
perf mem-events: Don't leak online CPU map
perf hist: In init, ensure mem_info is put on error paths
...
The NO_AUXTRACE build option was used when the __get_cpuid feature
test failed or if it was provided on the command line. The option no
longer avoids a dependency on a library and so having the option is
just adding complexity to the code base. Remove the option
CONFIG_AUXTRACE from Build files and HAVE_AUXTRACE_SUPPORT by assuming
it is always defined.
Signed-off-by: Ian Rogers <irogers@google.com>
Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Commit a808a2b35f ("tools build: Fix fixdep dependencies") broke the
perf build ("make -C tools/perf") by introducing two inadvertent
conflicts:
1) tools/build/Makefile includes tools/build/Makefile.include, which
defines a phony 'fixdep' target. This conflicts with the $(FIXDEP)
file target in tools/build/Makefile when OUTPUT is empty, causing
make to report duplicate recipes for the same target.
2) The FIXDEP variable in tools/build/Makefile conflicts with the
previously existing one in tools/perf/Makefile.perf.
Remove the unnecessary include of tools/build/Makefile.include from
tools/build/Makefile, and rename the FIXDEP variable in
tools/perf/Makefile.perf to FIXDEP_BUILT.
Fixes: a808a2b35f ("tools build: Fix fixdep dependencies")
Reported-by: Thorsten Leemhuis <linux@leemhuis.info>
Signed-off-by: Josh Poimboeuf <jpoimboe@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Thorsten Leemhuis <linux@leemhuis.info>
Link: https://patch.msgid.link/8881bc3321bd9fa58802e4f36286eefe3667806b.1760992391.git.jpoimboe@kernel.org
The jevents command expects all json files to be organized under a
single directory. When generating json files from scripts (to reduce
laborious copy and paste in the json) we don't want to generate the
json into the source directory if there is an OUTPUT directory
specified. This change adds a GEN_JSON for this case where the
GEN_JSON copies the JSON files to OUTPUT, only when OUTPUT is
specified. The Makefile.perf clean code is updated to clean up this
directory when present.
This patch is part of:
https://lore.kernel.org/lkml/20240926173554.404411-12-irogers@google.com/
which was similarly adding support for generating json in scripts for
the consumption of jevents.py.
Tested-by: Thomas Richter <tmricht@linux.ibm.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: James Clark <james.clark@linaro.org>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Since commit 22f72088ff ("tools headers: Update the syscall table with
the kernel sources") the arm64 syscall header is generated at build
time. Later, commit bfb713ea53 ("perf tools: Fix arm64 build by
generating unistd_64.h") added a dependency to libperf to guarantee that
this header was created before building libperf or perf itself.
However, libjvmti also requires this header but does not depend on
libperf, leading to build failures such as:
In file included from /usr/include/sys/syscall.h:24,
from /usr/include/syscall.h:1,
from jvmti/jvmti_agent.c:36:
tools/arch/arm64/include/uapi/asm/unistd.h:2:10: fatal error: asm/unistd_64.h: No such file or directory
2 | #include <asm/unistd_64.h>
Fix this by ensuring that libperf is built before libjvmti, so that
unistd_64.h is always available.
Fixes: 22f72088ff ("tools headers: Update the syscall table with the kernel sources")
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Vincent Minet <v.minet@criteo.com>
Link: https://lore.kernel.org/r/20250922053702.2688374-1-v.minet@criteo.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
If libperl is installed then the perf tool build will build against
it. There appears to be limited interest in the scripting support for
perl so let's make it opt-in and deprecate it.
With this patch applied you need to add LIBPERL=1 to get libperl
support in perf - there is no warning if libperl is missing, but
building will fail if libperl is missing and the build has LIBPERL=1.
The perf version output is changed to:
```
$ perf version --build-options
perf version 6.17.rc3.g8eca69269947
aio: [ on ] # HAVE_AIO_SUPPORT
bpf: [ on ] # HAVE_LIBBPF_SUPPORT
bpf_skeletons: [ on ] # HAVE_BPF_SKEL
debuginfod: [ on ] # HAVE_DEBUGINFOD_SUPPORT
dwarf: [ on ] # HAVE_LIBDW_SUPPORT
dwarf_getlocations: [ on ] # HAVE_LIBDW_SUPPORT
dwarf-unwind: [ on ] # HAVE_DWARF_UNWIND_SUPPORT
auxtrace: [ on ] # HAVE_AUXTRACE_SUPPORT
libbfd: [ OFF ] # HAVE_LIBBFD_SUPPORT ( tip: Deprecated, license incompatibility, use BUILD_NONDISTRO=1 and install binutils-dev[el] )
libbpf-strings: [ on ] # HAVE_LIBBPF_STRINGS_SUPPORT
libcapstone: [ on ] # HAVE_LIBCAPSTONE_SUPPORT
libdw-dwarf-unwind: [ on ] # HAVE_LIBDW_SUPPORT
libelf: [ on ] # HAVE_LIBELF_SUPPORT
libnuma: [ on ] # HAVE_LIBNUMA_SUPPORT
libopencsd: [ OFF ] # HAVE_CSTRACE_SUPPORT
libperl: [ OFF ] # HAVE_LIBPERL_SUPPORT ( tip:
Deprecated, use LIBPERL=1 and install libperl-dev to build with it )
libpfm4: [ on ] # HAVE_LIBPFM
libpython: [ on ] # HAVE_LIBPYTHON_SUPPORT
libslang: [ on ] # HAVE_SLANG_SUPPORT
libtraceevent: [ on ] # HAVE_LIBTRACEEVENT
libunwind: [ OFF ] # HAVE_LIBUNWIND_SUPPORT ( tip:
Deprecated, use LIBUNWIND=1 and install libunwind-dev[el] to build
with it )
lzma: [ on ] # HAVE_LZMA_SUPPORT
numa_num_possible_cpus: [ on ] # HAVE_LIBNUMA_SUPPORT
zlib: [ on ] # HAVE_ZLIB_SUPPORT
zstd: [ on ] # HAVE_ZSTD_SUPPORT
```
i.e. there is a tip saying about deprecation and how to get support
back.
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Charlie Jenkins <charlie@rivosinc.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Michael Petlan <mpetlan@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Quentin Monnet <qmo@kernel.org>
Cc: Steven Rostedt (VMware) <rostedt@goodmis.org>
Cc: Tomas Glozar <tglozar@redhat.com>
Cc: Yuzhuo Jing <yuzhuo@google.com>
Link: https://lore.kernel.org/lkml/aMrk03gigBlGcYLK@x1/
Link: https://lore.kernel.org/lkml/CAP-5=fVX+bLBRJCiziDi_hBySgv2NFtDoghtpheSSxVAvvETGw@mail.gmail.com
[ Keep the pre-existing perl-ExtUtils-Embed hint for Fedora/RHEL systems ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
When the clang toolchain has stack protection enabled, the bpf
skeletons build fails with:
error: A call to built-in function '__stack_chk_fail' is not supported.
Since stack-protector makes no sense for the BPF bits, just unconditionally
disable it.
See also similar case at 878625e1c7
Signed-off-by: Federico Pellegrin <fede@evolware.org>
Link: https://lore.kernel.org/r/20250718041224.12389-1-fede@evolware.org
[ rearrange long lines ]
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
When someone has a global shellcheckrc file, for example at
~/.config/shellcheckrc, with the directive 'shell=sh', building perf
will fail with many shellcheck errors like:
In tests/shell/base_probe/test_adding_kernel.sh line 294:
(( TEST_RESULT += $? ))
^---------------------^ SC3006 (warning): In POSIX sh, standalone ((..)) is undefined.
For more information:
https://www.shellcheck.net/wiki/SC3006 -- In POSIX sh, standalone ((..)) is...
make[5]: *** [tests/Build:91: tests/shell/base_probe/test_adding_kernel.sh.shellcheck_log] Error 1
Passing the '-s bash' option ensures that it runs correctly regardless
of a developers global configuration.
This patch adds '-s bash' and other options to the SHELLCHECK variable
in Makefile.perf and makes use of the variable consistently.
Signed-off-by: Collin Funk <collin.funk1@gmail.com>
Link: https://lore.kernel.org/r/63491dbc8439edf2e949d80e264b9d22332fea61.1751082075.git.collin.funk1@gmail.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
This is an end-to-end test for the PERF_RECORD_BPF_METADATA support.
It adds a new "bpf_metadata_perf_version" variable to perf's BPF programs,
so that when they are loaded, there will be at least one BPF program with
some metadata to parse. The test invokes "perf record" in a way that loads
one of those BPF programs, and then sifts through the output to find its
BPF metadata.
Signed-off-by: Blake Jones <blakejones@google.com>
Link: https://lore.kernel.org/r/20250612194939.162730-6-blakejones@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
When -s/--summary option is used, it doesn't need (augmented) arguments
of syscalls. Let's skip the augmentation and load another small BPF
program to collect the statistics in the kernel instead of copying the
data to the ring-buffer to calculate the stats in userspace. This will
be much more light-weight than the existing approach and remove any lost
events.
Let's add a new option --bpf-summary to control this behavior. I cannot
make it default because there's no way to get e_machine in the BPF which
is needed for detecting different ABIs like 32-bit compat mode.
No functional changes intended except for no more LOST events. :)
$ sudo ./perf trace -as --summary-mode=total --bpf-summary sleep 1
Summary of events:
total, 6194 events
syscall calls errors total min avg max stddev
(msec) (msec) (msec) (msec) (%)
--------------- -------- ------ -------- --------- --------- --------- ------
epoll_wait 561 0 4530.843 0.000 8.076 520.941 18.75%
futex 693 45 4317.231 0.000 6.230 500.077 21.98%
poll 300 0 1040.109 0.000 3.467 120.928 17.02%
clock_nanosleep 1 0 1000.172 1000.172 1000.172 1000.172 0.00%
ppoll 360 0 872.386 0.001 2.423 253.275 41.91%
epoll_pwait 14 0 384.349 0.001 27.453 380.002 98.79%
pselect6 14 0 108.130 7.198 7.724 8.206 0.85%
nanosleep 39 0 43.378 0.069 1.112 10.084 44.23%
...
Reviewed-by: Howard Chu <howardchu95@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Howard Chu <howardchu95@gmail.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/r/20250326044001.3503432-1-namhyung@kernel.org
[ Added fixup sent from Namhyung in response to my report to make it also dependent on CONFIG_TRACE ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Pull perf tools updates from Namhyung Kim:
"perf record:
- Introduce latency profiling using scheduler information.
The latency profiling is to show impacts on wall-time rather than
cpu-time. By tracking context switches, it can weight samples and
find which part of the code contributed more to the execution
latency.
The value (period) of the sample is weighted by dividing it by the
number of parallel execution at the moment. The parallelism is
tracked in perf report with sched-switch records. This will reduce
the portion that are run in parallel and in turn increase the
portion of serial executions.
For now, it's limited to profile processes, IOW system-wide
profiling is not supported. You can add --latency option to enable
this.
$ perf record --latency -- make -C tools/perf
I've run the above command for perf build which adds -j option to
make with the number of CPUs in the system internally. Normally
it'd show something like below:
$ perf report -F overhead,comm
...
#
# Overhead Command
# ........ ...............
#
78.97% cc1
6.54% python3
4.21% shellcheck
3.28% ld
1.80% as
1.37% cc1plus
0.80% sh
0.62% clang
0.56% gcc
0.44% perl
0.39% make
...
The cc1 takes around 80% of the overhead as it's the actual
compiler. However it runs in parallel so its contribution to
latency may be less than that. Now, perf report will show both
overhead and latency (if --latency was given at record time) like
below:
$ perf report -s comm
...
#
# Overhead Latency Command
# ........ ........ ...............
#
78.97% 48.66% cc1
6.54% 25.68% python3
4.21% 0.39% shellcheck
3.28% 13.70% ld
1.80% 2.56% as
1.37% 3.08% cc1plus
0.80% 0.98% sh
0.62% 0.61% clang
0.56% 0.33% gcc
0.44% 1.71% perl
0.39% 0.83% make
...
You can see latency of cc1 goes down to around 50% and python3 and
ld contribute a lot more than their overhead. You can use --latency
option in perf report to get the same result but ordered by
latency.
$ perf report --latency -s comm
perf report:
- As a side effect of the latency profiling work, it adds a new
output field 'latency' and a sort key 'parallelism'. The below is a
result from my system with 64 CPUs. The build was well-parallelized
but contained some serial portions.
$ perf report -s parallelism
...
#
# Overhead Latency Parallelism
# ........ ........ ...........
#
16.95% 1.54% 62
13.38% 1.24% 61
12.50% 70.47% 1
11.81% 1.06% 63
7.59% 0.71% 60
4.33% 12.20% 2
3.41% 0.33% 59
2.05% 0.18% 64
1.75% 1.09% 9
1.64% 1.85% 5
...
- Support Feodra mini-debuginfo which is a LZMA compressed symbol
table inside ".gnu_debugdata" ELF section.
perf annotate:
- Add --code-with-type option to enable data-type profiling with the
usual annotate output.
Instead of focusing on data structure, it shows code annotation
together with data type it accesses in case the instruction refers
to a memory location (and it was able to resolve the target data
type). Currently it only works with --stdio.
$ perf annotate --stdio --code-with-type
...
Percent | Source code & Disassembly of vmlinux for cpu/mem-loads,ldlat=30/pp (18 samples, percent: local period)
----------------------------------------------------------------------------------------------------------------------
: 0 0xffffffff81050610 <__fdget>:
0.00 : ffffffff81050610: callq 0xffffffff81c01b80 <__fentry__> # data-type: (stack operation)
0.00 : ffffffff81050615: pushq %rbp # data-type: (stack operation)
0.00 : ffffffff81050616: movq %rsp, %rbp
0.00 : ffffffff81050619: pushq %r15 # data-type: (stack operation)
0.00 : ffffffff8105061b: pushq %r14 # data-type: (stack operation)
0.00 : ffffffff8105061d: pushq %rbx # data-type: (stack operation)
0.00 : ffffffff8105061e: subq $0x10, %rsp
0.00 : ffffffff81050622: movl %edi, %ebx
0.00 : ffffffff81050624: movq %gs:0x7efc4814(%rip), %rax # 0x14e40 <current_task> # data-type: struct task_struct* +0
0.00 : ffffffff8105062c: movq 0x8d0(%rax), %r14 # data-type: struct task_struct +0x8d0 (files)
0.00 : ffffffff81050633: movl (%r14), %eax # data-type: struct files_struct +0 (count.counter)
0.00 : ffffffff81050636: cmpl $0x1, %eax
0.00 : ffffffff81050639: je 0xffffffff810506a9 <__fdget+0x99>
0.00 : ffffffff8105063b: movq 0x20(%r14), %rcx # data-type: struct files_struct +0x20 (fdt)
0.00 : ffffffff8105063f: movl (%rcx), %eax # data-type: struct fdtable +0 (max_fds)
0.00 : ffffffff81050641: cmpl %ebx, %eax
0.00 : ffffffff81050643: jbe 0xffffffff810506ef <__fdget+0xdf>
0.00 : ffffffff81050649: movl %ebx, %r15d
5.56 : ffffffff8105064c: movq 0x8(%rcx), %rdx # data-type: struct fdtable +0x8 (fd)
...
The "# data-type:" part was added with this change. The first few
entries are not very interesting. But later you can it accesses a
couple of fields in the task_struct, files_struct and fdtable.
perf trace:
- Support syscall tracing for different ABI. For example it can trace
system calls for 32-bit applications on 64-bit kernel
transparently.
- Add --summary-mode=total option to show global syscall summary. The
default is 'thread' to show per-thread syscall summary.
Python support:
- Add more interfaces to 'perf' module to parse events, and config,
enable or disable the event list properly so that it can implement
basic functionalities purely in Python. There is an example code
for these new interfaces in python/tracepoint.py.
- Add mypy and pylint support to enable build time checking. Fix some
code based on the findings from these tools.
Internals:
- Introduce io_dir__readdir() API to make directory traveral (usually
for proc or sysfs) efficient with less memory footprint.
JSON vendor events:
- Add events and metrics for ARM Neoverse N3 and V3
- Update events and metrics on various Intel CPUs
- Add/update events for a number of SiFive processors"
* tag 'perf-tools-for-v6.15-2025-03-27' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools: (229 commits)
perf bpf-filter: Fix a parsing error with comma
perf report: Fix a memory leak for perf_env on AMD
perf trace: Fix wrong size to bpf_map__update_elem call
perf tools: annotate asm_pure_loop.S
perf python: Fix setup.py mypy errors
perf test: Address attr.py mypy error
perf build: Add pylint build tests
perf build: Add mypy build tests
perf build: Rename TEST_LOGS to SHELL_TEST_LOGS
tools/build: Don't pass test log files to linker
perf bench sched pipe: fix enforced blocking reads in worker_thread
perf tools: Fix is_compat_mode build break in ppc64
perf build: filter all combinations of -flto for libperl
perf vendor events arm64 AmpereOneX: Fix frontend_bound calculation
perf vendor events arm64: AmpereOne/AmpereOneX: Mark LD_RETIRED impacted by errata
perf trace: Fix evlist memory leak
perf trace: Fix BTF memory leak
perf trace: Make syscall table stable
perf syscalltbl: Mask off ABI type for MIPS system calls
perf build: Remove Makefile.syscalls
...
