| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Commit b1c5efbfd92e ("perf parse-events: Remove hard coded legacy hardware
and cache parsing") introduced a bypass to PMU filtering to prevent uncore
PMUs from being filtered out during event parsing, which was required for
resolving `duration_time` and `uncore_freq` when running with `--cputype`.
However, this bypass was active whenever `pmu_filter` was set, which also
incorrectly bypassed filtering for the `--pmu-filter` option.
Introduce a `cputype_filter` boolean flag in `parse_events_state` and
`parse_events_option_args` to distinguish filtering initiated by
`--cputype` from that initiated by `--pmu-filter`. Restrict the core-only
check in `parse_events__filter_pmu()` to when `cputype_filter` is true.
Fixes: b1c5efbfd92e ("perf parse-events: Remove hard coded legacy hardware and cache parsing")
Assisted-by: Antigravity:gemini-3.1-pro
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Now the evlist is reference counted, add reference count checking so
that gets and puts are paired and easy to debug. Reference count
checking is documented here:
https://perfwiki.github.io/main/reference-count-checking/
This large patch is adding accessors to evlist functions and switching
to their use. There was some minor renaming as evlist__mmap is now an
accessor to the mmap variable, and the original evlist__mmap is
renamed to evlist__do_mmap.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alice Rogers <alice.mei.rogers@gmail.com>
Cc: Dapeng Mi <dapeng1.mi@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
This a no-op for most of the perf tool. The reference count is set to
1 at allocation, the put will see the 1, decrement it and perform the
delete.
The purpose for adding the reference count is for the python code. Prior
to this change the python code would clone evlists, but this has issues
if events are opened, etc.
This change adds a reference count for the evlists and a later change
will add it to evsels. The combination is needed for the python code to
operate correctly (not hit asserts in the evsel clone), but the changes
are broken apart for the sake of smaller patches.
Assisted-by: Gemini:gemini-3.1-pro-preview
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alice Rogers <alice.mei.rogers@gmail.com>
Cc: Dapeng Mi <dapeng1.mi@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
For evlist of a certain event/metric, the HEAD should be the event
leader. In some scenarios where uncore_xxx_0 does not exist, the event
leader is not the first element after sorting. For example, on my test
machine uncore_iio_0 does not exist, the event leader is uncore_iio_2.
However, in `evlist__cmp`, it was reordered based on the PMU name, which
makes uncore_iio_1 the HEAD of evlist, breaking the following merge
logic in `evsel__merge_aliases`.
The patch adds a loop at the end of
`parse_events__sort_events_and_fix_groups` to make sure the first
wildcard match is the earliest entry in the list, updating pointers
accordingly without breaking reordering detection.
Tested on device lacks uncore_iio_0, and `perf test` looks good.
Signed-off-by: Chun-Tse Shao <ctshao@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
This patch adds a new --pmu-filter option to perf-stat command to allow
filtering events on specific PMUs. This is useful when there are
multiple PMUs with same type (e.g. hisi_sicl2_cpa0 and hisi_sicl0_cpa0).
[root@localhost tmp]# perf stat -M cpa_p0_avg_bw
Performance counter stats for 'system wide':
19,417,779,115 hisi_sicl0_cpa0/cpa_cycles/ # 0.00 cpa_p0_avg_bw
0 hisi_sicl0_cpa0/cpa_p0_wr_dat/
0 hisi_sicl0_cpa0/cpa_p0_rd_dat_64b/
0 hisi_sicl0_cpa0/cpa_p0_rd_dat_32b/
19,417,751,103 hisi_sicl10_cpa0/cpa_cycles/ # 0.00 cpa_p0_avg_bw
0 hisi_sicl10_cpa0/cpa_p0_wr_dat/
0 hisi_sicl10_cpa0/cpa_p0_rd_dat_64b/
0 hisi_sicl10_cpa0/cpa_p0_rd_dat_32b/
19,417,730,679 hisi_sicl2_cpa0/cpa_cycles/ # 0.31 cpa_p0_avg_bw
75,635,749 hisi_sicl2_cpa0/cpa_p0_wr_dat/
18,520,640 hisi_sicl2_cpa0/cpa_p0_rd_dat_64b/
0 hisi_sicl2_cpa0/cpa_p0_rd_dat_32b/
19,417,674,227 hisi_sicl8_cpa0/cpa_cycles/ # 0.00 cpa_p0_avg_bw
0 hisi_sicl8_cpa0/cpa_p0_wr_dat/
0 hisi_sicl8_cpa0/cpa_p0_rd_dat_64b/
0 hisi_sicl8_cpa0/cpa_p0_rd_dat_32b/
19.417734480 seconds time elapsed
[root@localhost tmp]# perf stat --pmu-filter hisi_sicl2_cpa0 -M cpa_p0_avg_bw
Performance counter stats for 'system wide':
6,234,093,559 cpa_cycles # 0.60 cpa_p0_avg_bw
50,548,465 cpa_p0_wr_dat
7,552,182 cpa_p0_rd_dat_64b
0 cpa_p0_rd_dat_32b
6.234139320 seconds time elapsed
Signed-off-by: Qinxin Xia <xiaqinxin@huawei.com>
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The "Read backward ring buffer" test crashes on big-endian (e.g. s390x)
due to a NULL dereference when the backward mmap path isn't enabled.
Reproducer:
# ./perf test -F 'Read backward ring buffer'
Segmentation fault (core dumped)
# uname -m
s390x
#
Root cause:
get_config_terms() stores into evsel_config_term::val.val (u64) while later
code reads boolean fields such as evsel_config_term::val.overwrite.
On big-endian the 1-byte boolean is left-aligned, so writing
evsel_config_term::val.val = 1 is read back as
evsel_config_term::val.overwrite = 0,
leaving backward mmap disabled and a NULL map being used.
Store values in the union member that matches the term type, e.g.:
/* for OVERWRITE */
new_term->val.overwrite = 1; /* not new_term->val.val = 1 */
to fix this. Improve add_config_term() and add two more parameters for
string and value. Function add_config_term() now creates a complete node
element of type evsel_config_term and handles all evsel_config_term::val
union members.
Impact:
Enables backward mmap on big-endian and prevents the crash.
No change on little-endian.
Output after:
# ./perf test -Fv 44
--- start ---
Using CPUID IBM,9175,705,ME1,3.8,002f
mmap size 1052672B
mmap size 8192B
---- end ----
44: Read backward ring buffer : Ok
#
Fixes: 159ca97cd97c ("perf parse-events: Refactor get_config_terms() to remove macros")
Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Reviewed-by: Jan Polensky <japo@linux.ibm.com>
Reviewed-by: James Clark <james.clark@linaro.org>
Acked-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Tool events like duration_time don't need a perf_cpu_map that contains
all online CPUs.
Having such a perf_cpu_map causes overheads when iterating between
events for CPU affinity.
During parsing mark events that just read on a single CPU map index as
such, then during map propagation set up the evsel's CPUs and thereby
the evlists accordingly.
The setting cannot be done early in parsing as user CPUs are only fully
known when evlist__create_maps is called.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Andres Freund <andres@anarazel.de>
Cc: Dapeng Mi <dapeng1.mi@linux.intel.com>
Cc: Dr. David Alan Gilbert <linux@treblig.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Cc: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
This reverts commit d8d8a0b3603a9a8fa207cf9e4f292e81dc5d1008.
The setting of a user CPU map can cause an empty intersection when
combined with CPU 0 and the event removed. This later triggers a segv in
the stat-shadow logic. Let's put back a full online CPU map for now by
reverting this patch.
Closes: https://lore.kernel.org/linux-perf-users/cgja46br2smmznxs7kbeabs6zgv3b4olfqgh2fdp5mxk2yom4v@w6jjgov6hdi6/
Reported-by: Andres Freund <andres@anarazel.de>
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Dapeng Mi <dapeng1.mi@linux.intel.com>
Cc: Dr. David Alan Gilbert <linux@treblig.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Cc: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Merge with upstream to pick up fixes from perf-tools and from other
tools/ parts that interact with tools/perf.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
evlist__uniquify_evsel_names() only gets called in __parse_events() if
verbose is > 0. This means that the auto added "slots" events stay as
"slots" rather than being expanded to "cpu_core/slots/" unless Perf is
run in verbose mode. This is invisible to users when running Perf stat
because evlist__print_counters() always calls it regardless of verbose
mode before displaying.
The only thing this seems to affect is the test "Parsing of all PMU
events from sysfs" which fails when not run in verbose mode.
test__checkevent_pmu_events() always expects event names to be prefixed
with the pmu name, but this only happens for "slots" events after
evlist__uniquify_evsel_names() is called.
One fix could be to relax the test to accept the non prefixed name in
normal mode. But seeing as Perf stat uniquifies unconditionally, make
parse_events() do the same.
This fixes the following test failure:
$ perf test "Parsing of all PMU events from sysfs"
5.2: Parsing of all PMU events from sysfs : FAILED!
$
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: James Clark <james.clark@linaro.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Requiring the 'pmu->perf_event_attr_init_default' callback to be set to
track user changes is a bit of a trap to fall in. It's hard to see that
this is required when depending on the user change tracking.
It's possible to want all 0 defaults so not set it, but at the same time
still do some programmatic setting of configs with
evsel__set_config_if_unset(). Also if a PMU reverts to 0 defaults and
deletes its existing callback, it will silently break existing uses of
evsel__set_config_if_unset().
One way to fix this would be to assert in evsel__set_config_if_unset()
if the changes weren't tracked, but that would be a possibly untested
runtime failure. Instead, always track it as it's harmless and
simplifies testing too.
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: James Clark <james.clark@linaro.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Suzuki Poulouse <suzuki.poulose@arm.com>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Currently we only track which bits were set by the user in attr->config.
But all configN fields should be treated equally as they can all have
default and user overridden values.
Track them all by making get_config_chgs() generic and calling it once
for each config value.
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: James Clark <james.clark@linaro.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Suzuki Poulouse <suzuki.poulose@arm.com>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
The ADD_CONFIG_TERM() macros build the __type argument out of a partial
EVSEL__CONFIG_TERM_x enum name. This means that they can't be called
from a function where __type is a variable and it's also impossible to
grep the codebase to find usages of these enums as they're never typed
in full.
Fix this by removing the macros and replacing them with an
add_config_term() function. It seems the main reason these existed in
the first place was to avoid type punning and to write to a specific
field in the union, but the same thing can be achieved with a single
write to a u64 'val' field.
Running the Perf tests with "-fsanitize=undefined -fno-sanitize-recover"
results in no new issues as a result of this change.
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: James Clark <james.clark@linaro.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Suzuki Poulouse <suzuki.poulose@arm.com>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
If evsel__new_idx() returns NULL, the function currently jumps to label
'out_err'. Here, references to `cpus` and `pmu_cpus` are dropped.
Also, resources held by evsel->name and evsel->metric_id are freed.
But if evsel__new_idx() returns NULL, it can lead to NULL pointer
dereference.
Fixes: cd63c22168257a0b ("perf parse-events: Minor __add_event refactoring")
Signed-off-by: Faisal Bukhari <faisalbukhari523@gmail.com>
Reviewed-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
perf_event_attr has gained a new field, config4, so add support for it
extending the existing configN support.
Reviewed-by: Leo Yan <leo.yan@arm.com>
Reviewed-by: Ian Rogers <irogers@google.com>
Tested-by: Leo Yan <leo.yan@arm.com>
Signed-off-by: James Clark <james.clark@linaro.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The user and system time events can record on different CPUs, but for
all other events a single CPU map of just CPU 0 makes sense. In
parse-events detect a tool PMU and then pass the perf_event_attr so
that the tool_pmu can return CPUs specific for the event. This avoids
a CPU map of all online CPUs being used for events like
duration_time. Avoiding this avoids the evlist CPUs containing CPUs
for which duration_time just gives 0. Minimizing the evlist CPUs can
remove unnecessary sched_setaffinity syscalls that delay metric
calculations.
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
If verbose is enabled and parse_event is called, typically by tests,
log failures.
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Events with an X modifier were reordered within a group, for example
slots was made the leader in:
```
$ perf record -e '{cpu/mem-stores/ppu,cpu/slots/uX}' -- sleep 1
```
Fix by making `dont_regroup` evsels always use their index for
sorting. Make the cur_leader, when fixing the groups, be that of
`dont_regroup` evsel so that the `dont_regroup` evsel doesn't become a
leader.
On a tigerlake this patch corrects this and meets expectations in:
```
$ perf stat -e '{cpu/mem-stores/,cpu/slots/uX}' -a -- sleep 0.1
Performance counter stats for 'system wide':
83,458,652 cpu/mem-stores/
2,720,854,880 cpu/slots/uX
0.103780587 seconds time elapsed
$ perf stat -e 'slots,slots:X' -a -- sleep 0.1
Performance counter stats for 'system wide':
732,042,247 slots (48.96%)
643,288,155 slots:X (51.04%)
0.102731018 seconds time elapsed
```
Closes: https://lore.kernel.org/lkml/18f20d38-070c-4e17-bc90-cf7102e1e53d@linux.intel.com/
Fixes: 035c17893082 ("perf parse-events: Add 'X' modifier to exclude an event from being regrouped")
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Now that legacy hardware and cache events are in json, having the
lexer match the specific event is no longer necessary and generic PMU
parsing can take place. Because of this remove the specific term
parsing, event adding, and passing of alternate_hw_config which was
now always PERF_COUNT_HW_MAX.
This mirrors a similar change for software events in commit 6e9fa4131abb
("perf parse-events: Remove non-json software events").
With no hard coded legacy hardware or cache events the wild card, case
insensitivity, etc. is consistent for events. This does, however, mean
events like cycles will wild card against all PMUs. A change does the
same was originally posted and merged from:
https://lore.kernel.org/r/20240416061533.921723-10-irogers@google.com
and reverted by Linus in commit 4f1b067359ac ("Revert "perf
parse-events: Prefer sysfs/JSON hardware events over legacy"") due to
his dislike for the cycles behavior on ARM. Earlier patches in this
series make perf record event opening failures non-fatal and hide the
cycles event's failure to open on ARM in perf record, so it is
expected the behavior will now be transparent in perf record. perf
stat with a cycles event will wildcard open the event on all PMUs. As
cycles is a "default event", the perf stat behavior for default events
was updated to only open them on core/software PMUs.
The change to support legacy events with PMUs was done to clean up
Intel's hybrid PMU implementation. Having sysfs/json events with
increased priority to legacy was requested by Mark Rutland
<mark.rutland@arm.com> to fix Apple-M PMU issues wrt broken legacy
events on that PMU. It was requested that RISC-V be able to add events
to the perf tool json so the PMU driver didn't need to map legacy
events to config encodings:
https://lore.kernel.org/lkml/20240217005738.3744121-1-atishp@rivosinc.com/
A previous series of patches decreasing legacy hardware event
priorities was posted in:
https://lore.kernel.org/lkml/20250416045117.876775-1-irogers@google.com/
Namhyung Kim <namhyung@kernel.org> mentioned that hardware and
software events can be implemented similarly:
https://lore.kernel.org/lkml/aIJmJns2lopxf3EK@google.com/
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>
|
|
Now legacy hardware events are in json there's no need for a specific
printing routine that previously served for both hardware and software
events. The associated event_symbols_hw is also removed. To support
the previous filtered version use an event glob of "legacy hardware"
which matches the topic of the json events.
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>
|
|
Add the PMU terms legacy-hardware-config and
legacy-cache-config. These terms are similar to the config term in
that their values are assigned to the perf_event_attr config
value. They differ in that the PMU type is switched to be either
PERF_TYPE_HARDWARE or PERF_TYPE_HW_CACHE, and the PMU type is moved
into the extended type information of the config value. This will
allow later patches to add legacy events to json.
An example use of the terms is in the following:
```
$ perf stat -vv -e 'cpu/legacy-hardware-config=1/,cpu/legacy-cache-config=0x10001/' true
Using CPUID GenuineIntel-6-8D-1
Attempt to add: cpu/legacy-hardware-config=0x1/
..after resolving event: cpu/legacy-hardware-config=0x1/
Attempt to add: cpu/legacy-cache-config=0x10001/
..after resolving event: cpu/legacy-cache-config=0x10001/
Control descriptor is not initialized
------------------------------------------------------------
perf_event_attr:
type 0 (PERF_TYPE_HARDWARE)
size 136
config 0x1 (PERF_COUNT_HW_INSTRUCTIONS)
sample_type IDENTIFIER
read_format TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING
disabled 1
inherit 1
enable_on_exec 1
------------------------------------------------------------
sys_perf_event_open: pid 994937 cpu -1 group_fd -1 flags 0x8 = 3
------------------------------------------------------------
perf_event_attr:
type 3 (PERF_TYPE_HW_CACHE)
size 136
config 0x10001 (PERF_COUNT_HW_CACHE_RESULT_MISS | PERF_COUNT_HW_CACHE_OP_READ | PERF_COUNT_HW_CACHE_L1I)
sample_type IDENTIFIER
read_format TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING
disabled 1
inherit 1
enable_on_exec 1
------------------------------------------------------------
sys_perf_event_open: pid 994937 cpu -1 group_fd -1 flags 0x8 = 4
cpu/legacy-hardware-config=1/: -1: 1364046 414756 414756
cpu/legacy-cache-config=0x10001/: -1: 57453 414756 414756
cpu/legacy-hardware-config=1/: 1364046 414756 414756
cpu/legacy-cache-config=0x10001/: 57453 414756 414756
Performance counter stats for 'true':
1,364,046 cpu/legacy-hardware-config=1/
57,453 cpu/legacy-cache-config=0x10001/
0.001988593 seconds time elapsed
0.002194000 seconds user
0.000000000 seconds sys
```
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>
|
|
Now the events file isn't directly parsed from a FILE but stored in a
string prior to parsing, remove the FILE argument to the associated
scanner functions as they only ever pass NULL.
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>
|
|
When an event/alias is created for a PMU the terms are eagerly parsed
using parse_events_terms. For a command like perf stat or perf record,
the particular event/alias will be found, the terms parsed, the
terms cloned for use in the event parsing, and then the terms used to
configure the perf_event_attr. Events/aliases may be eagerly loaded,
such as from sysfs or in perf list, in which case the aliases terms
will be little or never used. To avoid redundant work, to avoid
cloning, and to reduce memory overhead, hold the terms for an event as
a string until they need handling as a term list. This may introduce
duplicate parsing if an event is repeated in a list, but this
situation is expected to be uncommon.
Measuring the number of instructions before and after with a sysfs
event and perf stat, there is a minor reduction in the number of
instructions executed by 0.3%.
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>
|
|
The term list when adding an event to a PMU is expected to have the
event name for the alias lookup. Also, set found_supported so that
-EINVAL isn't returned.
Fixes: 62593394f66a ("perf parse-events: Legacy cache names on all
PMUs and lower priority")
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>
|
|
Provide ratio-to-prev term which allows the user to
set the event sample period of two events corresponding
to a desired ratio.
If using on an Intel x86 platform with Auto Counter Reload support, also
set corresponding event's config2 attribute with a bitmask which
counters to reset and which counters to sample if the desired ratio is
met or exceeded.
On other platforms, only the sample period is affected by the
ratio-to-prev term.
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Thomas Falcon <thomas.falcon@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: Dapeng Mi <dapeng1.mi@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
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>
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 "Parsing of PMU event table metrics with fake PMUs" will test
metrics on machines/models that may be missing a PMU, in such a case
the fake_pmu should be used to avoid errors.
Metrics that get the cpumask from a different PMU, such as
"tsc/cpu=cpu_atom/", also need to be resilient in this test.
The parse_events_state fake_pmu is set when missing PMUs should be
ignored.
So that it can be queried, pass it to the config term functions, as well
as to get_config_cpu, then ignore failures when fake_pmu is set.
Some minor code refactoring to cut down on the indent and remove some
redundant checks.
Fixes: bd741d80dc65922c ("perf parse-events: Allow the cpu term to be a PMU or CPU range")
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.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: linux-actions@lists.infradead.org
Cc: Manivannan Sadhasivam <mani@kernel.org>
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: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250818190416.145274-2-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Start the creation of a tracepoint PMU abstraction. Tracepoint events
don't follow the regular sysfs perf conventions. Eventually the new
PMU abstraction will bridge the gap so tracepoint events look more
like regular perf ones.
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Link: https://lore.kernel.org/r/20250725185202.68671-5-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Remove the hard coded encodings from parse-events. This has the
consequence that software events are matched using the sysfs/json
priority, will be case insensitive and will be wildcarded across PMUs.
As there were software and hardware types in the parsing code, the
removal means software vs hardware logic can be removed and hardware
assumed.
Now the perf json provides detailed descriptions of software events,
remove the previous listing support that didn't contain event
descriptions. When globbing is required for the "sw" option in perf
list, use string PMU globbing as was done previously for the tool PMU.
The output of `perf list sw` command changed like this.
Before:
List of pre-defined events (to be used in -e or -M):
alignment-faults [Software event]
bpf-output [Software event]
cgroup-switches [Software event]
context-switches OR cs [Software event]
cpu-clock [Software event]
cpu-migrations OR migrations [Software event]
dummy [Software event]
emulation-faults [Software event]
major-faults [Software event]
minor-faults [Software event]
page-faults OR faults [Software event]
task-clock [Software event]
After:
List of pre-defined events (to be used in -e or -M):
software:
alignment-faults
[Number of kernel handled memory alignment faults. Unit: software]
bpf-output
[An event used by BPF programs to write to the perf ring buffer. Unit: software]
cgroup-switches
[Number of context switches to a task in a different cgroup. Unit: software]
context-switches
[Number of context switches [This event is an alias of cs]. Unit: software]
cpu-clock
[Per-CPU high-resolution timer based event. Unit: software]
cpu-migrations
[Number of times a process has migrated to a new CPU [This event is an alias of migrations]. Unit: software]
cs
[Number of context switches [This event is an alias of context-switches]. Unit: software]
dummy
[A placeholder event that doesn't count anything. Unit: software]
...
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Link: https://lore.kernel.org/r/20250725185202.68671-4-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Topdown metric events require grouping with a slots event. In perf
metrics this is currently achieved by metrics adding an unnecessary
"0 * tma_info_thread_slots". New TMA metrics trigger optimizations of
the metric expression that removes the event and breaks the metric due
to the missing but required event. Add a pass immediately before
sorting and fixing parsed events, that insert a slots event if one is
missing. Update test expectations to match this.
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250719030517.1990983-15-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Counting events system-wide with a specified CPU prior to this change
worked:
```
$ perf stat -e 'msr/tsc/,msr/tsc,cpu=cpu_core/,msr/tsc,cpu=cpu_atom/' -a sleep 1
Performance counter stats for 'system wide':
59,393,419,099 msr/tsc/
33,927,965,927 msr/tsc,cpu=cpu_core/
25,465,608,044 msr/tsc,cpu=cpu_atom/
```
However, when counting with process the counts became system wide:
```
$ perf stat -e 'msr/tsc/,msr/tsc,cpu=cpu_core/,msr/tsc,cpu=cpu_atom/' perf test -F 10
10.1: Basic parsing test : Ok
10.2: Parsing without PMU name : Ok
10.3: Parsing with PMU name : Ok
Performance counter stats for 'perf test -F 10':
59,233,549 msr/tsc/
59,227,556 msr/tsc,cpu=cpu_core/
59,224,053 msr/tsc,cpu=cpu_atom/
```
Make the handling of CPU maps with event parsing clearer. When an
event is parsed creating an evsel the cpus should be either the PMU's
cpumask or user specified CPUs.
Update perf_evlist__propagate_maps so that it doesn't clobber the user
specified CPUs. Try to make the behavior clearer, firstly fix up
missing cpumasks. Next, perform sanity checks and adjustments from the
global evlist CPU requests and for the PMU including simplifying to
the "any CPU"(-1) value. Finally remove the event if the cpumask is
empty.
So that events are opened with a CPU and a thread change stat's
create_perf_stat_counter to give both.
With the change things are fixed:
```
$ perf stat --no-scale -e 'msr/tsc/,msr/tsc,cpu=cpu_core/,msr/tsc,cpu=cpu_atom/' perf test -F 10
10.1: Basic parsing test : Ok
10.2: Parsing without PMU name : Ok
10.3: Parsing with PMU name : Ok
Performance counter stats for 'perf test -F 10':
63,704,975 msr/tsc/
47,060,704 msr/tsc,cpu=cpu_core/ (4.62%)
16,640,591 msr/tsc,cpu=cpu_atom/ (2.18%)
```
However, note the "--no-scale" option is used. This is necessary as
the running time for the event on the counter isn't the same as the
enabled time because the thread doesn't necessarily run on the CPUs
specified for the counter. All counter values are scaled with:
scaled_value = value * time_enabled / time_running
and so without --no-scale the scaled_value becomes very large. This
problem already exists on hybrid systems for the same reason. Here are
2 runs of the same code with an instructions event that counts the
same on both types of core, there is no real multiplexing happening on
the event:
```
$ perf stat -e instructions perf test -F 10
...
Performance counter stats for 'perf test -F 10':
87,896,447 cpu_atom/instructions/ (14.37%)
98,171,964 cpu_core/instructions/ (85.63%)
...
$ perf stat --no-scale -e instructions perf test -F 10
...
Performance counter stats for 'perf test -F 10':
13,069,890 cpu_atom/instructions/ (19.32%)
83,460,274 cpu_core/instructions/ (80.68%)
...
```
The scaling has inflated per-PMU instruction counts and the overall
count by 2x.
To fix this the kernel needs changing when a task+CPU event (or just
task event on hybrid) is scheduled out. A fix could be that the state
isn't inactive but off for such events, so that time_enabled counts
don't accumulate on them.
Reviewed-by: Thomas Falcon <thomas.falcon@intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250719030517.1990983-13-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Rename cpu_list to user_cpus. If a PMU isn't given, find it early from
the perf_event_attr. Make the pmu_cpus more explicitly a copy from the
PMU (except when user_cpus are given). Derive the cpus from pmu_cpus
and user_cpus as appropriate. Handle strdup errors on name and
metric_id.
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/20250719030517.1990983-11-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
own_cpus is generally the cpumask from the PMU. Rename to pmu_cpus to
try to make this clearer. Variable rename with no other changes.
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/20250719030517.1990983-7-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
On hybrid systems, events like msr/tsc/ will aggregate counts across
all CPUs. Often metrics only want a value like msr/tsc/ for the cores
on which the metric is being computed. Listing each CPU with terms
cpu=0,cpu=1.. is laborious and would need to be encoded for all
variations of a CPU model.
Allow the cpumask from a PMU to be an argument to the cpu term. For
example in the following the cpumask of the cstate_pkg PMU selects the
CPUs to count msr/tsc/ counter upon:
```
$ cat /sys/bus/event_source/devices/cstate_pkg/cpumask
0
$ perf stat -A -e 'msr/tsc,cpu=cstate_pkg/' -a sleep 0.1
Performance counter stats for 'system wide':
CPU0 252,621,253 msr/tsc,cpu=cstate_pkg/
0.101184092 seconds time elapsed
```
As the cpu term is now also allowed to be a string, allow it to encode
a range of CPUs (a list can't be supported as ',' is already a special
token).
The "event qualifiers" section of the `perf list` man page is updated
to detail the additional behavior. The man page formatting is tidied
up in this section, as it was incorrectly appearing within the
"parameterized events" section.
Reviewed-by: Thomas Falcon <thomas.falcon@intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250719030517.1990983-5-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>
|
|
Add missing breakpoint and raw types. Avoid a switch, just use a
lookup array. Switch the type to unsigned to avoid checking negative
values.
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250710235126.1086011-3-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Add perf_pmus__scan_for_event that only reads sysfs for pmus that
could contain a given event.
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250624231837.179536-2-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Commit 7b100989b4f6bce7 ("perf evlist: Remove __evlist__add_default")
changed to use "cycles:P" as a default event. But the problem is it
cannot set other default modifiers correctly.
perf kvm needs to set attr.exclude_host by default but it didn't work
because of the logic in the parse_events__modifier_list(). Also the
exclude_GH_default was applied only if ":u" modifier was specified -
which is strange. Move it out after handling the ":GH" and check
perf_host and perf_guest properly.
Before:
$ ./perf kvm record -vv true |& grep exclude
(nothing)
But specifying an event (without a modifier) works:
$ ./perf kvm record -vv -e cycles true |& grep exclude
exclude_host 1
After:
It now works for the both cases:
$ ./perf kvm record -vv true |& grep exclude
exclude_host 1
$ ./perf kvm record -vv -e cycles true |& grep exclude
exclude_host 1
Reviewed-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250606225431.2109754-1-namhyung@kernel.org
Fixes: 35c8d21371e9b342 ("perf tools: Don't set attr.exclude_guest by default")
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Add parse_uid_filter filter as a helper to parse_filter, that
constructs a uid filter string. As uid filters don't work with
tracepoint filters, add a is_possible_tp_filter function so the
tracepoint filter isn't attempted for tracepoint evsels.
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250604174545.2853620-4-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Rather than manually scanning PMUs, use evsel__find_pmu that can use
the PMU set during event parsing.
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250604174545.2853620-2-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The merge stat code fails for uncore events if they are repeated twice,
for example `perf stat -e clockticks,clockticks -I 1000` as the counts
of the second set of uncore events will be merged into the first
counter.
Reimplement the logic to have a first_wildcard_match so that merged
later events correctly merge into the first wildcard event that they
will be aggregated into.
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Chun-Tse Shao <ctshao@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Dr. David Alan Gilbert <linux@treblig.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: Levi Yun <yeoreum.yun@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250513215401.2315949-3-ctshao@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
'perf stat' has different uniquification logic to 'perf record' and perf
top. In the case of perf record and 'perf top' all hybrid event names
are uniquified.
'perf stat' is more disciplined respecting name config terms, libpfm4
events, etc.
'perf stat' will uniquify hybrid events and the non-core PMU cases
shouldn't apply to perf record or 'perf top'.
For consistency, remove the uniquification for 'perf record' and 'perf
top' and reuse the 'perf stat' uniquification, making the code more
globally visible for this.
Fix the detection of cross-PMU for disabling uniquify by correctly
setting last_pmu.
When setting uniquify on an evsel, make sure the PMUs between the 2
considered events differ otherwise the uniquify isn't adding value.
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Chun-Tse Shao <ctshao@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Dr. David Alan Gilbert <linux@treblig.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: Levi Yun <yeoreum.yun@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250513215401.2315949-2-ctshao@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
The -C option allows the CPUs for a list of events to be specified but
its not possible to set the CPU for a single event. Add a term to
allow this. The term isn't a general CPU list due to ',' already being
a special character in event parsing instead multiple cpu= terms may
be provided and they will be merged/unioned together.
An example of mixing different types of events counted on different CPUs:
```
$ perf stat -A -C 0,4-5,8 -e "instructions/cpu=0/,l1d-misses/cpu=4,cpu=5/,inst_retired.any/cpu=8/,cycles" -a sleep 0.1
Performance counter stats for 'system wide':
CPU0 6,979,225 instructions/cpu=0/ # 0.89 insn per cycle
CPU4 75,138 cpu/l1d-misses/
CPU5 1,418,939 cpu/l1d-misses/
CPU8 797,553 cpu/inst_retired.any,cpu=8/
CPU0 7,845,302 cycles
CPU4 6,546,859 cycles
CPU5 185,915,438 cycles
CPU8 2,065,668 cycles
0.112449242 seconds time elapsed
```
Committer testing:
root@number:~# grep -m1 "model name" /proc/cpuinfo
model name : AMD Ryzen 9 9950X3D 16-Core Processor
root@number:~# perf stat -A -e "instructions/cpu=0/,instructions,l1d-misses/cpu=4,cpu=5/,cycles" -a sleep 0.1
Performance counter stats for 'system wide':
CPU0 2,398,351 instructions/cpu=0/ # 0.44 insn per cycle
CPU0 2,398,152 instructions # 0.44 insn per cycle
CPU1 1,265,634 instructions # 0.49 insn per cycle
CPU2 606,087 instructions # 0.50 insn per cycle
CPU3 4,025,752 instructions # 0.52 insn per cycle
CPU4 4,236,810 instructions # 0.53 insn per cycle
CPU5 3,984,832 instructions # 0.66 insn per cycle
CPU6 434,132 instructions # 0.44 insn per cycle
CPU7 65,752 instructions # 0.41 insn per cycle
CPU8 459,083 instructions # 0.48 insn per cycle
CPU9 6,464,161 instructions # 1.31 insn per cycle
<SNIP>
root@number:~# perf stat -e "instructions/cpu=0/,instructions,l1d-misses/cpu=4,cpu=5/,cycles" -a sleep 0.
Performance counter stats for 'system wide':
144,822 instructions/cpu=0/ # 0.03 insn per cycle
4,666,114 instructions # 0.93 insn per cycle
2,583 l1d-misses
4,993,633 cycles
0.000868512 seconds time elapsed
root@number:~#
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Kan Liang <kan.liang@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: Dominique Martinet <asmadeus@codewreck.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Cc: Yicong Yang <yangyicong@hisilicon.com>
Link: https://lore.kernel.org/r/20250403194337.40202-5-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Also set the CPU map to all online CPU maps.
This is done so the behavior of legacy hardware and hardware cache
events better matches that of sysfs and JSON events during
__perf_evlist__propagate_maps().
Fix missing cpumap put in "Synthesize attr update" test.
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: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Dominique Martinet <asmadeus@codewreck.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Cc: Yicong Yang <yangyicong@hisilicon.com>
Link: https://lore.kernel.org/r/20250403194337.40202-4-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Add debug verbose output to show how evsels were reordered by
parse_events__sort_events_and_fix_groups(). For example:
```
$ perf record -v -e '{instructions,cycles}' true
Using CPUID GenuineIntel-6-B7-1
WARNING: events were regrouped to match PMUs
evlist after sorting/fixing: '{cpu_atom/instructions/,cpu_atom/cycles/},{cpu_core/instructions/,cpu_core/cycles/}'
```
Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Dominique Martinet <asmadeus@codewreck.org>
Cc: Dr. David Alan Gilbert <linux@treblig.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: Levi Yun <yeoreum.yun@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250402201549.4090305-6-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
The updated Intel vendor events add retirement latency for
graniterapids:
https://lore.kernel.org/lkml/20250322063403.364981-14-irogers@google.com/
This change makes those values available within an alias/event within a
PMU and saves them into the evsel at event parse time.
When no TPEBS data is available the default values are substituted in
for TMA metrics that are using retirement latency events - currently
just those on graniterapids.
Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-16-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
In the case of '{instructions,slots},faults,topdown-retiring' the
first event that must be grouped, slots, is ignored causing the
topdown-retiring event not to be adjacent to the group it needs to be
inserted into. Don't ignore the group members when computing the
force_grouped_index.
Make the force_grouped_index be for the leader of the group it is
within and always use it first rather than a group leader index so
that topdown events may be sorted from one group into another.
As the PMU name comparison applies to moving events in the same group
ensure the name ordering is always respected.
Change the group splitting logic to not group if there are no other
topdown events and to fix cases where the force group leader wasn't
being grouped with the other members of its group.
Reported-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Closes: https://lore.kernel.org/lkml/20250224083306.71813-2-dapeng1.mi@linux.intel.com/
Closes: https://lore.kernel.org/lkml/f7e4f7e8-748c-4ec7-9088-0e844392c11a@linux.intel.com/
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Link: https://lore.kernel.org/r/20250307023906.1135613-3-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Avoid DIR allocations when scanning sysfs by using io_dir for the
readdir implementation, that allocates about 1kb on the stack.
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250222061015.303622-7-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Wildcard PMU naming will match a name like pmu_1 to a PMU name like
pmu_10 but not to a PMU name like pmu_2 as the suffix forms part of
the match. No suffix matching will match pmu_10 to either pmu_1 or
pmu_2. Add or rename matching functions on PMU to make it clearer what
kind of matching is being performed.
Signed-off-by: Ian Rogers <irogers@google.com>
Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Link: https://lore.kernel.org/r/20250201074320.746259-4-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Add a new common config term "aux-action" to use for configuring AUX area
trace pause / resume. The value is a string that will be parsed in a
subsequent patch.
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20241216070244.14450-3-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
