For the sake of Intel topdown events commit 9eac5612da ("perf
stat: Don't skip failing group events") changed 'perf stat' error
handling making it so that more errors were fatal and didn't report
"<not supported>" events. The change outside of topdown events was
unintentional.
The notion of "fatal" error handling was introduced in commit
e0e6a6ca3a ("perf stat: Factor out open error handling") and
refined in commits like commit cb5ef60067 ("perf stat: Error out
unsupported group leader immediately") to be an approach for avoiding
later assertion failures in the code base.
This change fixes those issues and removes the notion of a fatal error
on an event. If all events fail to open then a fatal error occurs with
the previous fatal error message. This seems to best match the notion of
supported events and allowing some errors not to stop 'perf stat', while
allowing the truly fatal no event case to terminate the tool early.
The evsel->errored flag is only used in the stat code but always just
meaning !evsel->supported although there is a comment about it being
sticky. Force all evsels to be supported in evsel__init and then clear
this when evsel__open fails. When an event is tried the supported is
set to true again. This simplifies the notion of whether an evsel is
broken.
In the get_group_fd code, fail to get a group fd when the evsel isn't
supported. If the leader isn't supported then it is also expected that
there is no group_fd as the leader will have been skipped. Therefore
change the BUG_ON test to be on supported rather than skippable. This
corrects the assertion errors that were the reason for the previous
fatal error handling.
Fixes: 9eac5612da ("perf stat: Don't skip failing group events")
Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Chun-Tse Shao <ctshao@google.com>
Cc: Dapeng Mi <dapeng1.mi@linux.intel.com>
Cc: Howard Chu <howardchu95@gmail.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20251002220727.1889799-2-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The function parse_events__sort_events_and_fix_groups is needed to fix
uncore events like:
```
$ perf stat -e '{data_read,data_write}' ...
```
so that the multiple uncore PMUs have a group each of data_read and
data_write events.
The same function will perform architecture sorting and group fixing,
in particular for Intel topdown/perf-metric events. Grouping multiple
perf metric events together causes perf_event_open to fail as the
group can only support one. This means command lines like:
```
$ perf stat -e 'slots,slots' ...
```
fail as the slots events are forced into a group together to try to
satisfy the perf-metric event constraints.
As the user may know better than
parse_events__sort_events_and_fix_groups add a 'X' modifier to skip
its regrouping behavior. This allows the following to succeed rather
than fail on the second slots event being opened:
```
$ perf stat -e 'slots,slots:X' -a sleep 1
Performance counter stats for 'system wide':
6,834,154,071 cpu_core/slots/ (50.13%)
5,548,629,453 cpu_core/slots/X (49.87%)
1.002634606 seconds time elapsed
```
Closes: https://lore.kernel.org/lkml/20250822082233.1850417-1-dapeng1.mi@linux.intel.com/
Reported-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Reported-by: Xudong Hao <xudong.hao@intel.com>
Reviewed-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Howard Chu <howardchu95@gmail.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: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Cc: Yoshihiro Furudera <fj5100bi@fujitsu.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The session holds a perf_env pointer env. In UI code container_of is
used to turn the env to a session, but this assumes the session
header's env is in use. Rather than a dubious container_of, hold the
session in the evlist and derive the env from the session with
evsel__env, perf_session__env, etc.
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250724163302.596743-11-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Add evsel__open_per_cpu_and_thread that combines the operation of
evsel__open_per_cpu and evsel__open_per_thread so that an event
without the "any" cpumask can be opened with its cpumask and with
threads it specifies. Change the implementation of evsel__open_per_cpu
and evsel__open_per_thread to use evsel__open_per_cpu_and_thread to
make the implementation of those functions clearer.
Reviewed-by: Thomas Falcon <thomas.falcon@intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: James Clark <james.clark@linaro.org>
Link: https://lore.kernel.org/r/20250719030517.1990983-12-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Factor requested CPU warning out of evlist and into evsel. At the end
of adding an event, perform the warning check. To avoid repeatedly
testing if the cpu_list is empty, add a local variable.
```
$ perf stat -e cpu_atom/cycles,cpu=1/ -a true
WARNING: A requested CPU in '1' is not supported by PMU 'cpu_atom' (CPUs 16-27) for event 'cpu_atom/cycles/'
Performance counter stats for 'system wide':
<not supported> cpu_atom/cycles/
0.000781511 seconds time elapsed
```
Reviewed-by: Thomas Falcon <thomas.falcon@intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: James Clark <james.clark@linaro.org>
Link: https://lore.kernel.org/r/20250719030517.1990983-2-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
The existing logic would disable uniquification on an evlist or enable
it per evsel, this is unfortunate as uniquification is most needed
when events have the same name and so the whole evlist must be
considered. Change the initial disable uniquify on an evlist
processing to also set a needs_uniquify flag, for cases like the
matching event names. This must be done as an initial pass as
uniquification of an event name will change the behavior of the
check. Keep the per counter uniquification but now only uniquify event
names when the needs_uniquify flag is set.
Before this change a hwmon like temp1 wouldn't be uniquified and
afterwards it will (ie the PMU is added to the temp1 event's name).
Signed-off-by: Ian Rogers <irogers@google.com>
Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Link: https://lore.kernel.org/r/20250201074320.746259-6-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
The evsel__detect_missing_features() is to check if the attributes of
the evsel is supported or not. But it checks the attribute based on the
given evsel, it might miss something if the attr doesn't have the bit or
give incorrect results if the event is special.
Also it maintains the order of the feature that was added to the kernel
which means it can assume older features should be supported once it
detects the current feature is working. To minimized the confusion and
to accurately check the kernel features, I think it's better to use a
software event and go through all the features at once.
Also make the function static since it's only used in evsel.c.
Reviewed-by: James Clark <james.clark@linaro.org>
Reviewed-by: Ravi Bangoria <ravi.bangoria@amd.com>
Acked-by: Kan Liang <kan.liang@linux.intel.com>
Cc: James Clark <james.clark@arm.com>
Cc: Atish Patra <atishp@atishpatra.org>
Cc: Mingwei Zhang <mizhang@google.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Cc: Palmer Dabbelt <palmer@rivosinc.com>
Link: https://lore.kernel.org/r/20241016062359.264929-6-namhyung@kernel.org
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Rather than treat tool events as a special kind of event, create a
tool only PMU where the events/aliases match the existing
duration_time, user_time and system_time events. Remove special
parsing and printing support for the tool events, but add function
calls for when PMU functions are called on a tool_pmu.
Move the tool PMU code in evsel into tool_pmu.c to better encapsulate
the tool event behavior in that file.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20241002032016.333748-5-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
The "perf record" tool will now default to this new mode if the user
specifies a sampling group when not in system-wide mode, and when
"--no-inherit" is not specified.
This change updates evsel to allow the combination of inherit
and PERF_SAMPLE_READ.
A fallback is implemented for kernel versions where this feature is not
supported.
Signed-off-by: Ben Gainey <ben.gainey@arm.com>
Cc: james.clark@arm.com
Link: https://lore.kernel.org/r/20241001121505.1009685-3-ben.gainey@arm.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
The branch counters logging (A.K.A LBR event logging) introduces a
per-counter indication of precise event occurrences in LBRs. The kernel
only dumps the number of occurrences into a record. The perf tool has
to map the number to the corresponding event.
Add evlist__update_br_cntr() to go through the evlist to pick the
events that are configured to be logged. Assign a logical idx to track
them, and add the total number of the events in the leader event.
The total number will be used to allocate the space to save the branch
counters for a block. The logical idx will be used to locate the
corresponding event quickly in the following patches.
It only needs to iterate the evlist once. The
evsel__has_branch_counters() is also optimized.
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: https://lore.kernel.org/r/20240813160208.2493643-4-kan.liang@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
When retire_latency value is used in a metric formula, evsel would fork
a 'perf record' process with "-e" and "-W" options. 'perf record' will
collect required retire_latency values in parallel while 'perf stat' is
collecting counting values.
At the point of time that 'perf stat' stops counting, evsel would stop
'perf record' by sending sigterm signal to 'perf record' process.
Sampled data will be processed to get retire latency value. Another
thread is required to synchronize between 'perf stat' and 'perf record'
when we pass data through pipe.
Retire_latency evsel is not opened for 'perf stat' so that there is no
counter wasted on it. This commit includes code suggested by Namhyung to
adjust reading size for groups that include retire_latency evsels.
In current :R parsing implementation, the parser would recognize events
with retire_latency modifier and insert them into the evlist like a
normal event. Ideally, we need to avoid counting these events.
In this commit, at the time when a retire_latency evsel is read, set the
retire latency value processed from the sampled data to count value.
This sampled retire latency value will be used for metric calculation
and final event count print out. No special metric calculation and event
print out code required for retire_latency events.
Reviewed-by: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Samantha Alt <samantha.alt@intel.com>
Link: https://lore.kernel.org/r/20240720062102.444578-4-weilin.wang@intel.com
[ Squashed the 3rd and 4th commit in the series to keep it building patch by patch ]
[ Constified the 'struct perf_tool' pointer in process_sample_event() ]
[ Use perf_tool__init(&tool, false) to address a segfault I reported and Ian/Weilin diagnosed ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tool events unnecessarily open a dummy perf event which is useless
even with `perf record` which will still open a dummy event. Change
the behavior of tool events so:
- duration_time - call `rdclock` on open and then report the count as
a delta since the start in evsel__read_counter. This moves code out
of builtin-stat making it more general purpose.
- user_time/system_time - open the fd as either `/proc/pid/stat` or
`/proc/stat` for cases like system wide. evsel__read_counter will
read the appropriate field out of the procfs file. These values
were previously supplied by wait4, if the procfs read fails then
the wait4 values are used, assuming the process/thread terminated.
By reading user_time and system_time this way, interval mode, per
PID and per CPU can be supported although there are restrictions
given what the files provide (e.g. per PID can't be combined with
per CPU).
Opening any of the tool events for `perf record` is changed to return
invalid.
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Cc: James Clark <james.clark@arm.com>
Cc: Dmitrii Dolgov <9erthalion6@gmail.com>
Cc: Ze Gao <zegao2021@gmail.com>
Cc: Song Liu <song@kernel.org>
Cc: Leo Yan <leo.yan@linux.dev>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20240503232849.17752-1-irogers@google.com
Add a new branch filter, "counter", for the branch counter option. It is
used to mark the events which should be logged in the branch. If it is
applied with the -j option, the counters of all the events should be
logged in the branch. If the legacy kernel doesn't support the new
branch sample type, switching off the branch counter filter.
The stored counter values in each branch are displayed right after the
regular branch stack information via perf report -D.
Usage examples:
# perf record -e "{branch-instructions,branch-misses}:S" -j any,counter
Only the first event, branch-instructions, collect the LBR. Both
branch-instructions and branch-misses are marked as logged events. The
occurrences information of them can be found in the branch stack
extension space of each branch.
# perf record -e "{cpu/branch-instructions,branch_type=any/,cpu/branch-misses,branch_type=counter/}"
Only the first event, branch-instructions, collect the LBR. Only the
branch-misses event is marked as a logged event.
Committer notes:
I noticed 'perf test "Sample parsing"' failing, reported to the list and
Kan provided a patch that checks if the evsel has a leader and that
evsel->evlist is set, the comment in the source code further explains
it.
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexey Bayduraev <alexey.v.bayduraev@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Tinghao Zhang <tinghao.zhang@intel.com>
Link: https://lore.kernel.org/r/20231025201626.3000228-8-kan.liang@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
evsel__increase_rlimit() helper does nothing with evsel, and description
of the functionality is inaccurate, rename it and move to util/rlimit.c.
By the way, fix a checkppatch warning about misplaced license tag:
WARNING: Misplaced SPDX-License-Identifier tag - use line 1 instead
#160: FILE: tools/perf/util/rlimit.h:3:
/* SPDX-License-Identifier: LGPL-2.1 */
No functional change.
Signed-off-by: Yang Jihong <yangjihong1@huawei.com>
Link: https://lore.kernel.org/r/20231023033144.1011896-1-yangjihong1@huawei.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
In the default mode, the current output of the metricgroup include both
events and metrics, which is not necessary and just makes the output
hard to read. Since different ARCHs (even different generations in the
same ARCH) may use different events. The output also vary on different
platforms.
For a metricgroup, only outputting the value of each metric is good
enough.
Add a new field default_metricgroup in evsel to indicate an event of the
default metricgroup. For those events, printout() should print the
metricgroup name rather than each event.
Add perf_stat__skip_metric_event() to skip the evsel in the Default
metricgroup, if it's not running or not the metric event.
Add print_metricgroup_header_t to pass the functions which print the
display name of each metricgroup in the Default metricgroup. Support all
three output methods.
Factor out perf_stat__print_shadow_stats_metricgroup() to print out each
metrics.
On SPR:
Before:
./perf_old stat sleep 1
Performance counter stats for 'sleep 1':
0.54 msec task-clock:u # 0.001 CPUs utilized
0 context-switches:u # 0.000 /sec
0 cpu-migrations:u # 0.000 /sec
68 page-faults:u # 125.445 K/sec
540,970 cycles:u # 0.998 GHz
556,325 instructions:u # 1.03 insn per cycle
123,602 branches:u # 228.018 M/sec
6,889 branch-misses:u # 5.57% of all branches
3,245,820 TOPDOWN.SLOTS:u # 18.4 % tma_backend_bound
# 17.2 % tma_retiring
# 23.1 % tma_bad_speculation
# 41.4 % tma_frontend_bound
564,859 topdown-retiring:u
1,370,999 topdown-fe-bound:u
603,271 topdown-be-bound:u
744,874 topdown-bad-spec:u
12,661 INT_MISC.UOP_DROPPING:u # 23.357 M/sec
1.001798215 seconds time elapsed
0.000193000 seconds user
0.001700000 seconds sys
After:
$ ./perf stat sleep 1
Performance counter stats for 'sleep 1':
0.51 msec task-clock:u # 0.001 CPUs utilized
0 context-switches:u # 0.000 /sec
0 cpu-migrations:u # 0.000 /sec
68 page-faults:u # 132.683 K/sec
545,228 cycles:u # 1.064 GHz
555,509 instructions:u # 1.02 insn per cycle
123,574 branches:u # 241.120 M/sec
6,957 branch-misses:u # 5.63% of all branches
TopdownL1 # 17.5 % tma_backend_bound
# 22.6 % tma_bad_speculation
# 42.7 % tma_frontend_bound
# 17.1 % tma_retiring
TopdownL2 # 21.8 % tma_branch_mispredicts
# 11.5 % tma_core_bound
# 13.4 % tma_fetch_bandwidth
# 29.3 % tma_fetch_latency
# 2.7 % tma_heavy_operations
# 14.5 % tma_light_operations
# 0.8 % tma_machine_clears
# 6.1 % tma_memory_bound
1.001712086 seconds time elapsed
0.000151000 seconds user
0.001618000 seconds sys
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ahmad Yasin <ahmad.yasin@intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: https://lore.kernel.org/r/20230616031420.3751973-3-kan.liang@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
perf tools fixes for v6.4: 2nd batch
- Fix BPF CO-RE naming convention for checking the availability of fields on
'union perf_mem_data_src' on the running kernel.
- Remove the use of llvm-strip on BPF skel object files, not needed, fixes a
build breakage when the llvm package, that contains it in most distros, isn't
installed.
- Fix tools that use both evsel->{bpf_counter_list,bpf_filters}, removing them from a
union.
- Remove extra "--" from the 'perf ftrace latency' --use-nsec option,
previously it was working only when using the '-n' alternative.
- Don't stop building when both binutils-devel and a C++ compiler isn't
available to compile the alternative C++ demangle support code, disable that
feature instead.
- Sync the linux/in.h and coresight-pmu.h header copies with the kernel sources.
- Fix relative include path to cs-etm.h.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
'struct evsel' uses a union for the two lists. This turned out to be
error prone.
For example:
[root@quaco ~]# perf stat --bpf-prog 5246
Error: cpu-clock event does not have sample flags 66c660
failed to set filter "(null)" on event cpu-clock with 2 (No such file or directory)
[root@quaco ~]# perf stat --bpf-prog 5246
Fix this issue by separating the two lists.
Fixes: 56ec9457a4 ("perf bpf filter: Implement event sample filtering")
Signed-off-by: Song Liu <song@kernel.org>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: kernel-team@meta.com
Link: https://lore.kernel.org/r/20230519235757.3613719-1-song@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
'perf stat' with no arguments will use default events and metrics. These
events may fail to open even with kernel and hypervisor disabled. When
these fail then the permissions error appears even though they were
implicitly selected. This is particularly a problem with the automatic
selection of the TopdownL1 metric group on certain architectures like
Skylake:
$ perf stat true
Error:
Access to performance monitoring and observability operations is limited.
Consider adjusting /proc/sys/kernel/perf_event_paranoid setting to open
access to performance monitoring and observability operations for processes
without CAP_PERFMON, CAP_SYS_PTRACE or CAP_SYS_ADMIN Linux capability.
More information can be found at 'Perf events and tool security' document:
https://www.kernel.org/doc/html/latest/admin-guide/perf-security.html
perf_event_paranoid setting is 2:
-1: Allow use of (almost) all events by all users
Ignore mlock limit after perf_event_mlock_kb without CAP_IPC_LOCK
>= 0: Disallow raw and ftrace function tracepoint access
>= 1: Disallow CPU event access
>= 2: Disallow kernel profiling
To make the adjusted perf_event_paranoid setting permanent preserve it
in /etc/sysctl.conf (e.g. kernel.perf_event_paranoid = <setting>)
$
This patch adds skippable evsels that when they fail to open won't cause
termination and will appear as "<not supported>" in output. The
TopdownL1 events, from the metric group, are marked as skippable. This
turns the failure above to:
$ perf stat perf bench internals synthesize
Computing performance of single threaded perf event synthesis by
synthesizing events on the perf process itself:
Average synthesis took: 49.287 usec (+- 0.083 usec)
Average num. events: 3.000 (+- 0.000)
Average time per event 16.429 usec
Average data synthesis took: 49.641 usec (+- 0.085 usec)
Average num. events: 11.000 (+- 0.000)
Average time per event 4.513 usec
Performance counter stats for 'perf bench internals synthesize':
1,222.38 msec task-clock:u # 0.993 CPUs utilized
0 context-switches:u # 0.000 /sec
0 cpu-migrations:u # 0.000 /sec
162 page-faults:u # 132.529 /sec
774,445,184 cycles:u # 0.634 GHz (49.61%)
1,640,969,811 instructions:u # 2.12 insn per cycle (59.67%)
302,052,148 branches:u # 247.102 M/sec (59.69%)
1,807,718 branch-misses:u # 0.60% of all branches (59.68%)
5,218,927 CPU_CLK_UNHALTED.REF_XCLK:u # 4.269 M/sec
# 17.3 % tma_frontend_bound
# 56.4 % tma_retiring
# nan % tma_backend_bound
# nan % tma_bad_speculation (60.01%)
536,580,469 IDQ_UOPS_NOT_DELIVERED.CORE:u # 438.965 M/sec (60.33%)
<not supported> INT_MISC.RECOVERY_CYCLES_ANY:u
5,223,936 CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE:u # 4.274 M/sec (40.31%)
774,127,250 CPU_CLK_UNHALTED.THREAD:u # 633.297 M/sec (50.34%)
1,746,579,518 UOPS_RETIRED.RETIRE_SLOTS:u # 1.429 G/sec (50.12%)
1,940,625,702 UOPS_ISSUED.ANY:u # 1.588 G/sec (49.70%)
1.231055525 seconds time elapsed
0.258327000 seconds user
0.965749000 seconds sys
$
The event INT_MISC.RECOVERY_CYCLES_ANY:u is skipped as it can't be
opened with paranoia 2 on Skylake. With a lower paranoia, or as root,
all events/metrics are computed.
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Kan Liang <kan.liang@linux.intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ahmad Yasin <ahmad.yasin@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Edward Baker <edward.baker@intel.com>
Cc: Florian Fischer <florian.fischer@muhq.space>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@arm.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kang Minchul <tegongkang@gmail.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Cc: Rob Herring <robh@kernel.org>
Cc: Samantha Alt <samantha.alt@intel.com>
Cc: Stephane Eranian <eranian@google.com>
Cc: Sumanth Korikkar <sumanthk@linux.ibm.com>
Cc: Suzuki Poulouse <suzuki.poulose@arm.com>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Cc: Tiezhu Yang <yangtiezhu@loongson.cn>
Cc: Weilin Wang <weilin.wang@intel.com>
Cc: Xing Zhengjun <zhengjun.xing@linux.intel.com>
Cc: Yang Jihong <yangjihong1@huawei.com>
Link: https://lore.kernel.org/r/20230502223851.2234828-3-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
It seems that perf stat -b <prog id> doesn't produce any results:
$ perf stat -e cycles -b 4 -I 10000 -vvv
Control descriptor is not initialized
cycles: 0 0 0
time counts unit events
10.007641640 <not supported> cycles
Looks like this happens because fentry/fexit progs are getting loaded, but the
corresponding perf event is not enabled and not added into the events bpf map.
I think there is some mixing up between two type of bpf support, one for bperf
and one for bpf_profiler. Both are identified via evsel__is_bpf, based on which
perf events are enabled, but for the latter (bpf_profiler) a perf event is
required. Using evsel__is_bperf to check only bperf produces expected results:
$ perf stat -e cycles -b 4 -I 10000 -vvv
Control descriptor is not initialized
------------------------------------------------------------
perf_event_attr:
size 136
sample_type IDENTIFIER
read_format TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING
disabled 1
exclude_guest 1
------------------------------------------------------------
sys_perf_event_open: pid -1 cpu 0 group_fd -1 flags 0x8 = 3
------------------------------------------------------------
[...perf_event_attr for other CPUs...]
------------------------------------------------------------
cycles: 309426 169009 169009
time counts unit events
10.010091271 309426 cycles
The final numbers correspond (at least in the level of magnitude) to the
same metric obtained via bpftool.
Fixes: 112cb56164 ("perf stat: Introduce config stat.bpf-counter-events")
Reviewed-by: Song Liu <song@kernel.org>
Signed-off-by: Dmitrii Dolgov <9erthalion6@gmail.com>
Tested-by: Song Liu <song@kernel.org>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/r/20230412182316.11628-1-9erthalion6@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Thomas Falcon
Ian Rogers
Namhyung Kim
Howard Chu
Arnaldo Carvalho de Melo
Adrian Hunter
Ben Gainey
Kan Liang
Weilin Wang
Dominique Martinet
Ze Gao
Yang Jihong
Ravi Bangoria
Song Liu
Dmitrii Dolgov
James Clark