| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The struct dump_regs contains 512 bytes of cache_regs, meaning the two
values in perf_sample contribute 1088 bytes of its total 1384 bytes
size. Initializing this much memory has a cost reported by Tavian
Barnes <tavianator@tavianator.com> as about 2.5% when running `perf
script --itrace=i0`:
https://lore.kernel.org/lkml/d841b97b3ad2ca8bcab07e4293375fb7c32dfce7.1736618095.git.tavianator@tavianator.com/
Adrian Hunter <adrian.hunter@intel.com> replied that the zero
initialization was necessary and couldn't simply be removed.
This patch aims to strike a middle ground of still zeroing the
perf_sample, but removing 79% of its size by make user_regs and
intr_regs optional pointers to zalloc-ed memory. To support the
allocation accessors are created for user_regs and intr_regs. To
support correct cleanup perf_sample__init and perf_sample__exit
functions are created and added throughout the code base.
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250113194345.1537821-1-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
I found that it failed to load a binary using --symfs option. Say I
have a binary in /home/user/prog/xxx and a perf data file with it. If I
move them to a different machine and use --symfs, it tries to find the
binary in some locations under symfs using dso__read_binary_type_filename(),
but not the last one.
${symfs}/usr/lib/debug/home/user/prog/xxx.debug
${symfs}/usr/lib/debug/home/user/prog/xxx
${symfs}/home/user/prog/.debug/xxx
/home/user/prog/xxx
It should check ${symfs}/home/usr/prog/xxx. Let's fix it.
Reviewed-by: Ian Rogers <irogers@google.com>
Acked-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Link: https://lore.kernel.org/r/20250212221445.437481-1-namhyung@kernel.org
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
It was only used in perf trace and it switched to use hashmap instead.
Let's delete the code.
Reviewed-by: Howard Chu <howardchu95@gmail.com>
Acked-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Link: https://lore.kernel.org/r/20250205205443.1986408-4-namhyung@kernel.org
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Some topdown related metrics may fail on hybrid machines.
$ perf stat -M tma_frontend_bound
Cannot resolve IDs for tma_frontend_bound:
cpu_atom@TOPDOWN_FE_BOUND.ALL@ / (8 * cpu_atom@CPU_CLK_UNHALTED.CORE@)
In the find_tool_events(), the tool_pmu__event_to_str() is used to
compare the tool_events. It only checks the event name, no PMU or arch.
So the tool_events[TOOL_PMU__EVENT_SLOTS] is set to true, because the
p-core Topdown metrics has "slots" event.
The tool_events is shared. So when parsing the e-core metrics, the
"slots" is automatically added.
The "slots" event as a tool event should only be available on arm64. It
has a different meaning on X86. The tool_pmu__skip_event() intends
handle the case. Apply it for tool_pmu__event_to_str() as well.
There is a lack of sanity check in the expr__get_id(). Add the check.
Closes: https://lore.kernel.org/lkml/608077bc-4139-4a97-8dc4-7997177d95c4@linux.intel.com/
Fixes: 069057239a67 ("perf tool_pmu: Move expr literals to tool_pmu")
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Reviewed-by: Ian Rogers <irogers@google.com>
Cc: thomas.falcon@intel.com
Link: https://lore.kernel.org/r/20250207152844.302167-1-kan.liang@linux.intel.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The last use of machine__fprintf_vmlinux_path() was removed in 2011 by
commit ab81f3fd350c ("perf top: Reuse the 'report' hist_entry/hists
classes")
mmap_cpu_mask__duplicate() was added in 2021 by
commit 6bd006c6eb7f ("perf mmap: Introduce mmap_cpu_mask__duplicate()")
but hasn't been used since.
Remove them.
Signed-off-by: Dr. David Alan Gilbert <linux@treblig.org>
Tested-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250204220545.456435-1-linux@treblig.org
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
To get the various fixes in the current master.
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>
|
|
Counter merging was added in commit 942c5593393d ("perf stat: Add
perf_stat_merge_counters()"), however, it merges events with the same
name on different PMUs regardless of whether the different PMUs are
actually of the same type (ie they differ only in the suffix on the
PMU). For hwmon events there may be a temp1 event on every PMU, but
the PMU names are all unique and don't differ just by a suffix. The
merging is over eager and will merge all the hwmon counters together
meaning an aggregated and very large temp1 value is shown. The same
would be true for say cache events and memory controller events where
the PMUs differ but the event names are the same.
Fix the problem by correctly saying two PMUs alias when they differ
only by suffix.
Note, there is an overlap with evsel's merged_stat with aggregation
and the evsel's metric_leader where aggregation happens for metrics.
Fixes: 942c5593393d ("perf stat: Add perf_stat_merge_counters()")
Signed-off-by: Ian Rogers <irogers@google.com>
Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Link: https://lore.kernel.org/r/20250201074320.746259-5-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>
|
|
Rather than scanning core or all PMUs, allow pmu_read_sysfs to read
some combination of core, other, hwmon and tool PMUs. The PMUs that
should be read and are already read are held as bitmaps. It is known
that a "hwmon_" prefix is necessary for a hwmon PMU's name, similarly
with "tool", so only scan those PMUs in situations the PMU name or the
PMU's type number make sense to.
The number of openat system calls reduces from 276 to 98 for a hwmon
event. The number of openats for regular perf events isn't changed.
Signed-off-by: Ian Rogers <irogers@google.com>
Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Link: https://lore.kernel.org/r/20250201074320.746259-3-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
evsel__is_hybrid returns true if there are multiple core PMUs and the
evsel is for a core PMU. Determining the number of core PMUs can
require loading/scanning PMUs. There's no point doing the scanning if
evsel for the is_hybrid test isn't core so reorder the tests to reduce
PMU scanning.
Signed-off-by: Ian Rogers <irogers@google.com>
Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Link: https://lore.kernel.org/r/20250201074320.746259-2-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
perf test 11 hwmon fails on s390 with this error
# ./perf test -Fv 11
--- start ---
---- end ----
11.1: Basic parsing test : Ok
--- start ---
Testing 'temp_test_hwmon_event1'
Using CPUID IBM,3931,704,A01,3.7,002f
temp_test_hwmon_event1 -> hwmon_a_test_hwmon_pmu/temp_test_hwmon_event1/
FAILED tests/hwmon_pmu.c:189 Unexpected config for
'temp_test_hwmon_event1', 292470092988416 != 655361
---- end ----
11.2: Parsing without PMU name : FAILED!
--- start ---
Testing 'hwmon_a_test_hwmon_pmu/temp_test_hwmon_event1/'
FAILED tests/hwmon_pmu.c:189 Unexpected config for
'hwmon_a_test_hwmon_pmu/temp_test_hwmon_event1/',
292470092988416 != 655361
---- end ----
11.3: Parsing with PMU name : FAILED!
#
The root cause is in member test_event::config which is initialized
to 0xA0001 or 655361. During event parsing a long list event parsing
functions are called and end up with this gdb call stack:
#0 hwmon_pmu__config_term (hwm=0x168dfd0, attr=0x3ffffff5ee8,
term=0x168db60, err=0x3ffffff81c8) at util/hwmon_pmu.c:623
#1 hwmon_pmu__config_terms (pmu=0x168dfd0, attr=0x3ffffff5ee8,
terms=0x3ffffff5ea8, err=0x3ffffff81c8) at util/hwmon_pmu.c:662
#2 0x00000000012f870c in perf_pmu__config_terms (pmu=0x168dfd0,
attr=0x3ffffff5ee8, terms=0x3ffffff5ea8, zero=false,
apply_hardcoded=false, err=0x3ffffff81c8) at util/pmu.c:1519
#3 0x00000000012f88a4 in perf_pmu__config (pmu=0x168dfd0, attr=0x3ffffff5ee8,
head_terms=0x3ffffff5ea8, apply_hardcoded=false, err=0x3ffffff81c8)
at util/pmu.c:1545
#4 0x00000000012680c4 in parse_events_add_pmu (parse_state=0x3ffffff7fb8,
list=0x168dc00, pmu=0x168dfd0, const_parsed_terms=0x3ffffff6090,
auto_merge_stats=true, alternate_hw_config=10)
at util/parse-events.c:1508
#5 0x00000000012684c6 in parse_events_multi_pmu_add (parse_state=0x3ffffff7fb8,
event_name=0x168ec10 "temp_test_hwmon_event1", hw_config=10,
const_parsed_terms=0x0, listp=0x3ffffff6230, loc_=0x3ffffff70e0)
at util/parse-events.c:1592
#6 0x00000000012f0e4e in parse_events_parse (_parse_state=0x3ffffff7fb8,
scanner=0x16878c0) at util/parse-events.y:293
#7 0x00000000012695a0 in parse_events__scanner (str=0x3ffffff81d8
"temp_test_hwmon_event1", input=0x0, parse_state=0x3ffffff7fb8)
at util/parse-events.c:1867
#8 0x000000000126a1e8 in __parse_events (evlist=0x168b580,
str=0x3ffffff81d8 "temp_test_hwmon_event1", pmu_filter=0x0,
err=0x3ffffff81c8, fake_pmu=false, warn_if_reordered=true,
fake_tp=false) at util/parse-events.c:2136
#9 0x00000000011e36aa in parse_events (evlist=0x168b580,
str=0x3ffffff81d8 "temp_test_hwmon_event1", err=0x3ffffff81c8)
at /root/linux/tools/perf/util/parse-events.h:41
#10 0x00000000011e3e64 in do_test (i=0, with_pmu=false, with_alias=false)
at tests/hwmon_pmu.c:164
#11 0x00000000011e422c in test__hwmon_pmu (with_pmu=false)
at tests/hwmon_pmu.c:219
#12 0x00000000011e431c in test__hwmon_pmu_without_pmu (test=0x1610368
<suite.hwmon_pmu>, subtest=1) at tests/hwmon_pmu.c:23
where the attr::config is set to value 292470092988416 or 0x10a0000000000
in line 625 of file ./util/hwmon_pmu.c:
attr->config = key.type_and_num;
However member key::type_and_num is defined as union and bit field:
union hwmon_pmu_event_key {
long type_and_num;
struct {
int num :16;
enum hwmon_type type :8;
};
};
s390 is big endian and Intel is little endian architecture.
The events for the hwmon dummy pmu have num = 1 or num = 2 and
type is set to HWMON_TYPE_TEMP (which is 10).
On s390 this assignes member key::type_and_num the value of
0x10a0000000000 (which is 292470092988416) as shown in above
trace output.
Fix this and export the structure/union hwmon_pmu_event_key
so the test shares the same implementation as the event parsing
functions for union and bit fields. This should avoid
endianess issues on all platforms.
Output after:
# ./perf test -F 11
11.1: Basic parsing test : Ok
11.2: Parsing without PMU name : Ok
11.3: Parsing with PMU name : Ok
#
Fixes: 531ee0fd4836 ("perf test: Add hwmon "PMU" test")
Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Reviewed-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250131112400.568975-1-tmricht@linux.ibm.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
This is also used in util/comm.c now, so instead of selectively doing
the feature test, always do it. If it's ever used anywhere else it's
less likely to cause another build failure.
This doesn't remove the need to manually include libc_compat.h, and
missing that will still cause an error for glibc < 2.26. There isn't a
way to fix that without poisoning reallocarray like libbpf did, but that
has other downsides like making memory debugging tools less useful. So
for Perf keep it like this and we'll have to fix up any missed includes.
Fixes the following build error:
util/comm.c:152:31: error: implicit declaration of function
'reallocarray' [-Wimplicit-function-declaration]
152 | tmp = reallocarray(comm_strs->strs,
| ^~~~~~~~~~~~
Fixes: 13ca628716c6 ("perf comm: Add reference count checking to 'struct comm_str'")
Reported-by: Ali Utku Selen <ali.utku.selen@arm.com>
Signed-off-by: James Clark <james.clark@linaro.org>
Reviewed-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250129154405.777533-1-james.clark@linaro.org
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Legacy events typically don't have a PMU when added leading to
mismatched legacy/non-legacy cases in find_stat. Use evsel__find_pmu
to make sure the evsel PMU is looked up. Update the evsel__find_pmu
code to look for the PMU using the extended config type or, for legacy
hardware/hw_cache events on non-hybrid systems, just use the core PMU.
Before:
```
$ perf stat -e cycles,cpu/instructions/ -a sleep 1
Performance counter stats for 'system wide':
215,309,764 cycles
44,326,491 cpu/instructions/
1.002555314 seconds time elapsed
```
After:
```
$ perf stat -e cycles,cpu/instructions/ -a sleep 1
Performance counter stats for 'system wide':
990,676,332 cycles
1,235,762,487 cpu/instructions/ # 1.25 insn per cycle
1.002667198 seconds time elapsed
```
Fixes: 3612ca8e2935 ("perf stat: Fix the hard-coded metrics calculation on the hybrid")
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: James Clark <james.clark@linaro.org>
Tested-by: Leo Yan <leo.yan@arm.com>
Tested-by: Atish Patra <atishp@rivosinc.com>
Link: https://lore.kernel.org/r/20250109222109.567031-3-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Add helper to get the name of the evsel's PMU. This handles the case
where there's no sysfs PMU via parse_events event_type helper.
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: James Clark <james.clark@linaro.org>
Tested-by: Leo Yan <leo.yan@arm.com>
Tested-by: Atish Patra <atishp@rivosinc.com>
Link: https://lore.kernel.org/r/20250109222109.567031-2-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Now that filename__read_int() returns -errno instead of -1 these
statements need to be updated otherwise error values will be used as
die IDs.
This appears as a -2 die ID when the platform doesn't export one:
$ perf stat --per-core -a -- true
S36-D-2-C0 1 9.45 msec cpu-clock
And the session topology test fails:
$ perf test -vvv topology
CPU 0, core 0, socket 36
CPU 1, core 1, socket 36
CPU 2, core 2, socket 36
CPU 3, core 3, socket 36
FAILED tests/topology.c:137 Cpu map - Die ID doesn't match
---- end(-1) ----
38: Session topology : FAILED!
Fixes: 05be17eed774 ("tool api fs: Correctly encode errno for read/write open failures")
Reported-by: Thomas Richter <tmricht@linux.ibm.com>
Signed-off-by: James Clark <james.clark@linaro.org>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20241218115552.912517-1-james.clark@linaro.org
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Prior to this change a string was used which could cause issues with
an unrecognized disassembler in symbol__disassembler. Change to
initializing an array of perf_disassembler enum values. If a value
already exists then adding it a second time is ignored to avoid array
out of bounds problems present in the previous code, it also allows a
statically sized array and removes memory allocation needs. Errors in
the disassembler string are reported when the config is parsed during
perf annotate or perf top start up. If the array is uninitialized
after processing the config file the default llvm, capstone then
objdump values are added but without a need to parse a string.
Fixes: a6e8a58de629 ("perf disasm: Allow configuring what disassemblers to use")
Closes: https://lore.kernel.org/lkml/CAP-5=fUdfCyxmEiTpzS2uumUp3-SyQOseX2xZo81-dQtWXj6vA@mail.gmail.com/
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20250124043856.1177264-1-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
As reported by Namhyung Kim and acknowledged by Qiao Zhao (link:
https://lore.kernel.org/linux-perf-users/20241206001436.1947528-1-namhyung@kernel.org/),
on certain machines, perf trace failed to load the BPF program into the
kernel. The verifier runs perf trace's BPF program for up to 1 million
instructions, returning an E2BIG error, whereas the perf trace BPF
program should be much less complex than that. This patch aims to fix
the issue described above.
The E2BIG problem from clang-15 to clang-16 is cause by this line:
} else if (size < 0 && size >= -6) { /* buffer */
Specifically this check: size < 0. seems like clang generates a cool
optimization to this sign check that breaks things.
Making 'size' s64, and use
} else if ((int)size < 0 && size >= -6) { /* buffer */
Solves the problem. This is some Hogwarts magic.
And the unbounded access of clang-12 and clang-14 (clang-13 works this
time) is fixed by making variable 'aug_size' s64.
As for this:
-if (aug_size > TRACE_AUG_MAX_BUF)
- aug_size = TRACE_AUG_MAX_BUF;
+aug_size = args->args[index] > TRACE_AUG_MAX_BUF ? TRACE_AUG_MAX_BUF : args->args[index];
This makes the BPF skel generated by clang-18 work. Yes, new clangs
introduce problems too.
Sorry, I only know that it works, but I don't know how it works. I'm not
an expert in the BPF verifier. I really hope this is not a kernel
version issue, as that would make the test case (kernel_nr) *
(clang_nr), a true horror story. I will test it on more kernel versions
in the future.
Fixes: 395d38419f18: ("perf trace augmented_raw_syscalls: Add more check s to pass the verifier")
Reported-by: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Howard Chu <howardchu95@gmail.com>
Tested-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20241213023047.541218-1-howardchu95@gmail.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
An evsel idx may not be stable due to sorting, evlist removal,
etc. Try to reduce it being part of APIs by explicitly passing the
evsel in annotate code. Internally the code just reads evsel->core.idx
so behavior is unchanged.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Chen Ni <nichen@iscas.ac.cn>
Cc: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
Link: https://lore.kernel.org/r/20250117181848.690474-1-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
When a filtered column is not present in the sort order, profiles become
arbitrary broken. Filtered and non-filtered entries are collapsed
together, and the filtered-by field ends up with a random value (either
from a filtered or non-filtered entry). If we end up with filtered
entry/value, then the whole collapsed entry will be filtered out and will
be missing in the profile. If we end up with non-filtered entry/value,
then the overhead value will be wrongly larger (include some subset
of filtered out samples).
This leads to very confusing profiles. The problem is hard to notice,
and if noticed hard to understand. If the filter is for a single value,
then it can be fixed by adding the corresponding field to the sort order
(provided user understood the problem). But if the filter is for multiple
values, it's impossible to fix b/c there is no concept of binary sorting
based on filter predicate (we want to group all non-filtered values in
one bucket, and all filtered values in another).
Examples of affected commands:
perf report --tid=123
perf report --sort overhead,symbol --comm=foo,bar
Fix this by considering filtered status as the highest priority
sort/collapse predicate.
As a side effect this effectively adds a new feature of showing profile
where several lines are combined based on arbitrary filtering predicate.
For example, showing symbols from binaries foo and bar combined together,
but not from other binaries; or showing combined overhead of several
particular threads.
Signed-off-by: Dmitry Vyukov <dvyukov@google.com>
Link: https://lore.kernel.org/r/359dc444ce94d20e59d3a9e360c36fbeac833a04.1736927981.git.dvyukov@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Application of cmp/sort/collapse fmt callbacks is duplicated 6 times.
Factor it into a common helper function. NFC.
Signed-off-by: Dmitry Vyukov <dvyukov@google.com>
Link: https://lore.kernel.org/r/84c4b55614e24a344f86ae0db62e8fa8f251f874.1736927981.git.dvyukov@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
To allow for setting a variable from some other tool, like with the
"wallclock" patchset needs to allow the user to opt-in to having
that key in the sort order for 'perf report'.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Dmitriy Vyukov <dvyukov@google.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: James Clark <james.clark@linaro.org>
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: Stephane Eranian <eranian@google.com>
Link: https://lore.kernel.org/lkml/Z4akewi7UPXpagce@x1
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Rename err to out to avoid confusion because buf is still supposed to be
freed in non error cases.
Reviewed-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: James Clark <james.clark@linaro.org>
Tested-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Dr. David Alan Gilbert <linux@treblig.org>
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: Leo Yan <leo.yan@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20241211085525.519458-3-james.clark@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
When kernel is built without debuginfo, running 'perf record' with
--off-cpu results in segfault as below:
./perf record --off-cpu -e dummy sleep 1
libbpf: kernel BTF is missing at '/sys/kernel/btf/vmlinux', was CONFIG_DEBUG_INFO_BTF enabled?
libbpf: failed to find '.BTF' ELF section in /lib/modules/6.13.0-rc3+/build/vmlinux
libbpf: failed to find valid kernel BTF
Segmentation fault (core dumped)
The backtrace pointed to:
#0 0x00000000100fb17c in btf.type_cnt ()
#1 0x00000000100fc1a8 in btf_find_by_name_kind ()
#2 0x00000000100fc38c in btf.find_by_name_kind ()
#3 0x00000000102ee3ac in off_cpu_prepare ()
#4 0x000000001002f78c in cmd_record ()
#5 0x00000000100aee78 in run_builtin ()
#6 0x00000000100af3e4 in handle_internal_command ()
#7 0x000000001001004c in main ()
Code sequence is:
static void check_sched_switch_args(void)
{
struct btf *btf = btf__load_vmlinux_btf();
const struct btf_type *t1, *t2, *t3;
u32 type_id;
type_id = btf__find_by_name_kind(btf, "btf_trace_sched_switch",
BTF_KIND_TYPEDEF);
btf__load_vmlinux_btf() fails when CONFIG_DEBUG_INFO_BTF is not enabled.
Here bpf__find_by_name_kind() calls btf__type_cnt() with NULL btf value
and results in segfault.
To fix this, add a check to see if btf is not NULL before invoking
bpf__find_by_name_kind().
Reviewed-by: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Disha Goel <disgoel@linux.vnet.ibm.com>
Cc: Hari Bathini <hbathini@linux.ibm.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Madhavan Srinivasan <maddy@linux.ibm.com>
Link: https://lore.kernel.org/r/20241223135813.8175-1-atrajeev@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
In machine__create_module(), it reads /proc/modules to get a list of
modules in the system. The file shows the start address (of text) and
the size of the module so it uses the info to reconstruct system memory
maps for symbol resolution.
But module memory consists of multiple segments and they can be
scaterred. Currently perf tools assume they are contiguous and see some
overlaps. This can confuse the tool when it finds a map containing a
given address.
As we mostly care about the function symbols in the text segment, it can
fixup the size or end address of modules when there's an overlap. We
can use maps__fixup_end() which updates the end address using the start
address of the next map.
Ideally it should be able to track other segments (like data/rodata),
but that would require some changes in /proc/modules IMHO.
Reported-by: Blake Jones <blakejones@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Daniel Gomez <da.gomez@samsung.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Petr Pavlu <petr.pavlu@suse.com>
Cc: Sami Tolvanen <samitolvanen@google.com>
Link: https://lore.kernel.org/r/20241218220453.203069-1-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
When there are more than one symbols at the same address, it needs to
choose which one is better. In choose_best_symbol() it didn't check the
type of symbols. It's possible to have labels in other symbols and in
that case, it would be better to pick the actual symbol over the labels.
To minimize the possible impact on other symbols, I only check NOTYPE
symbols specifically.
$ readelf -sW vmlinux | grep -e __do_softirq -e __softirqentry_text_start
105089: ffffffff82000000 814 FUNC GLOBAL DEFAULT 1 __do_softirq
111954: ffffffff82000000 0 NOTYPE GLOBAL DEFAULT 1 __softirqentry_text_start
The commit 77b004f4c5c3c90b tried to do the same by not giving the size
to the label symbols but it seems there's some label-only symbols in asm
code. Let's restore the original code and choose the right symbol using
type of the symbols.
Fixes: 77b004f4c5c3c90b ("perf symbol: Do not fixup end address of labels")
Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Ian Rogers <irogers@google.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Link: http://lore.kernel.org/lkml/Z3b-DqBMnNb4ucEm@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
cxx_demangle_sym is weak in case demangle-cxx.c replaces the
definition in symbol-elf.c. When demangle-cxx.c is built
HAVE_CXA_DEMANGLE_SUPPORT is defined, as such the define can be used
to avoid a weak symbol.
As weak symbols are outside of the C standard their use can lead to
strange behaviors, in particular with LTO, as well as causing issues to
be hidden at link time.
Signed-off-by: Ian Rogers <irogers@google.com>
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: 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>
Link: https://lore.kernel.org/r/20241119031754.1021858-1-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
All architectures now support HAVE_SYSCALL_TABLE_SUPPORT, so the flag is
no longer needed. With the removal of the flag, the related
GENERIC_SYSCALL_TABLE can also be removed.
libaudit was only used as a fallback for when HAVE_SYSCALL_TABLE_SUPPORT
was not defined, so libaudit is also no longer needed for any
architecture.
Signed-off-by: Charlie Jenkins <charlie@rivosinc.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: Arnd Bergmann <arnd@arndb.de>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Guo Ren <guoren@kernel.org>
Cc: Günther Noack <gnoack@google.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mickaël Salaün <mic@digikod.net>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250108-perf_syscalltbl-v6-16-7543b5293098@rivosinc.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Use the generic scripts to generate headers from the syscall table
instead of the custom ones for s390.
Signed-off-by: Charlie Jenkins <charlie@rivosinc.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Guo Ren <guoren@kernel.org>
Cc: Günther Noack <gnoack@google.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mickaël Salaün <mic@digikod.net>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250108-perf_syscalltbl-v6-15-7543b5293098@rivosinc.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Use the generic scripts to generate headers from the syscall table
instead of the custom ones for powerpc.
Signed-off-by: Charlie Jenkins <charlie@rivosinc.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Guo Ren <guoren@kernel.org>
Cc: Günther Noack <gnoack@google.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mickaël Salaün <mic@digikod.net>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250108-perf_syscalltbl-v6-14-7543b5293098@rivosinc.com
Link: https://lore.kernel.org/lkml/20250110100505.78d81450@canb.auug.org.au
[ Stephen Rothwell noticed on linux-next that the powerpc build for perf was broken and ...]
Link: https://lore.kernel.org/lkml/20250109-perf_powerpc_spu-v1-1-c097fc43737e@rivosinc.com
[ ... Charlie fixed it up and asked for it to be squashed to avoid breaking bisection. ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Use the generic scripts to generate headers from the syscall table for
mips.
Signed-off-by: Charlie Jenkins <charlie@rivosinc.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Guo Ren <guoren@kernel.org>
Cc: Günther Noack <gnoack@google.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mickaël Salaün <mic@digikod.net>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250108-perf_syscalltbl-v6-13-7543b5293098@rivosinc.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
loongarch uses a syscall table, use that in perf instead of using unistd.h.
Signed-off-by: Charlie Jenkins <charlie@rivosinc.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Guo Ren <guoren@kernel.org>
Cc: Günther Noack <gnoack@google.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mickaël Salaün <mic@digikod.net>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250108-perf_syscalltbl-v6-12-7543b5293098@rivosinc.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
arm64 uses a syscall table, use that in perf instead of using unistd.h.
Signed-off-by: Charlie Jenkins <charlie@rivosinc.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Guo Ren <guoren@kernel.org>
Cc: Günther Noack <gnoack@google.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mickaël Salaün <mic@digikod.net>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250108-perf_syscalltbl-v6-11-7543b5293098@rivosinc.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Use the generic scripts to generate headers from the syscall table for
both 32- and 64-bit x86.
Signed-off-by: Charlie Jenkins <charlie@rivosinc.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: Arnd Bergmann <arnd@arndb.de>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Guo Ren <guoren@kernel.org>
Cc: Günther Noack <gnoack@google.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mickaël Salaün <mic@digikod.net>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250108-perf_syscalltbl-v6-8-7543b5293098@rivosinc.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Currently each architecture in perf independently generates syscall
headers.
Adapt the work that has gone into unifying syscall header
implementations in the kernel to work with perf tools.
Introduce this framework with riscv at first. riscv previously relied on
libaudit, but with this change, perf tools for riscv no longer needs
this external dependency.
Signed-off-by: Charlie Jenkins <charlie@rivosinc.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Guo Ren <guoren@kernel.org>
Cc: Günther Noack <gnoack@google.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Leo Yan <leo.yan@linux.dev>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mickaël Salaün <mic@digikod.net>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250108-perf_syscalltbl-v6-1-7543b5293098@rivosinc.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Remove unused symbol_conf.h include.
First, it's just unused. Second, it's problematic since this is a C++
file, and most perf headers don't compile as C++. So if any other
includes are added to symbol_conf.h, it may break the build.
Signed-off-by: Dmitriy Vyukov <dvyukov@google.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20250108070248.237943-1-dvyukov@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
When adding support for refconunt checking a cut'n'paste made this
function, that is just an accessor to a bool member of 'struct nsinfo',
return a pid_t, when that member is a boolean, fix it.
Fixes: bcaf0a97858de7ab ("perf namespaces: Add functions to access nsinfo")
Reported-by: Francesco Nigro <fnigro@redhat.com>
Reported-by: Ilan Green <igreen@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Clark Williams <williams@redhat.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Yonatan Goldschmidt <yonatan.goldschmidt@granulate.io>
Link: https://lore.kernel.org/r/20241206204828.507527-6-acme@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
not in the same pidns
When running 'perf record' outside a container and the java agent inside
a container the jit_repipe_code_load() and friends will emit
PERF_RECORD_MMAP2 entries for the jitdump records and will check if we
need to fixup the pid/tid:
nspid = jr->load.pid;
pid = jr_entry_pid(jd, jr);
tid = jr_entry_tid(jd, jr);
The jr_entry_pid() function looks if we're in the same pidns:
static pid_t jr_entry_pid(struct jit_buf_desc *jd, union jr_entry *jr)
{
if (jd->nsi && nsinfo__in_pidns(jd->nsi))
return nsinfo__tgid(jd->nsi);
return jr->load.pid;
}
But since the thread, populated from perf.data records, try to figure
out if in the same pidns by actually trying, on the system where 'perf
inject' is running to open a procfs file (a bug that remains to be
fixed), assuming that if it is not possible that is because that thread
terminated and thus we can't get its namespace info and tolerates
nsinfo__init() failing, noting only that that namespace can't be
entered, so don't even try.
But we can kinda get at least that info (thread->nsinfo->in_pidns) from
the data in the perf.data file, namely the pid and tid in the
PERF_RECORD_MMAP2 for the jit-<PID>.dump file generated from the java
agent, if the PERF_RECORD_MMAP2->pid is the same as what is in the
jitdump file, then we're in the same namespace, otherwise we need to use
the PERF_RECORD_MMAP2->pid.
This all has to be revamped for this jitdump + running perf from
outside, as the meaning of in_pidns is being abused, the initialization
of nsinfo->pid with the value coming from the PERF_RECORD_MMAP2 data is
wrong as it is the pid _outside_ the container since perf was running
there.
The hack in this patch at least produces the expected result in this
scenario by following the assumptions in the current codebase for
finding maps and for generating the PERF_RECORD_MMAP2 for the ELF files
synthesized from the jitdump records in jit_repipe_code_load(), etc.s
Reported-by: Francesco Nigro <fnigro@redhat.com>
Reported-by: Ilan Green <igreen@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Clark Williams <williams@redhat.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Yonatan Goldschmidt <yonatan.goldschmidt@granulate.io>
Link: https://lore.kernel.org/r/20241206204828.507527-5-acme@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
When we're processing a perf.data file we will, for every thread in that
file do a machine__findnew_thread(machine, pid, tid) that when that pid
is seen for the first time will create a 'struct thread' representing
it.
That in turn will call nsinfo__new() -> nsinfo__init() and there it will
assume we're running live, which is wrong and will need to be addressed
in a followup patch.
The nsinfo__new() assumes that if we can't access that thread it has
already finished and will ignore the -1 return from nsinfo__init(), just
taking notes to avoid trying to enter in that namespace, since it isn't
there anymore, a race.
When doing this from 'perf inject', tho, we can fill in parts of that
nsinfo from what we get from the PERF_RECORD_MMAP2 (pid, tid) and in the
jitdump file name, that has the form of jit-<PID>.dump.
So if the pid in the jitdump file name is not the one in the
PERF_RECORD_MMAP2, we can assume that its the pid of the process
_inside_ the namespace, and that perf was runing outside that namespace.
This will be done in the following patch.
Reported-by: Francesco Nigro <fnigro@redhat.com>
Reported-by: Ilan Green <igreen@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Clark Williams <williams@redhat.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Yonatan Goldschmidt <yonatan.goldschmidt@granulate.io>
Link: https://lore.kernel.org/r/20241206204828.507527-4-acme@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
When the java agent is running inside a container it will emit mmaps
with the format:
⬢ [acme@toolbox a]$ perf report -D | grep PERF_RECORD_MMAP | grep \.dump
0 0x15c400 [0x90]: PERF_RECORD_MMAP2 3308868/3308868: [0x7fb8de6cb000(0x1000) @ 0 08:14 3222905945 0]: r-xp /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jit-1.dump
⬢ [acme@toolbox a]$
Since perf is running from outside the container it sees the pid 3308868
in PERF_RECORD_MMAP2, while the agent saw the pid of the profiled app
inside the container, 1.
The previous validation was:
if (pid && pid2 != nsinfo__nstgid(nsi))
pid2 at this point is '1' (/jit-1.dump), so it considers this as a
malformed jitdump mmap and refuses to process it.
The test ends up as:
if (3308868 && 1 != 3308868)
which is true and the jitdump is not processed.
Since 1 in the container namespace is really 3308868 in the namespace
that perf is running, consider this a valid mmap.
We need to make perf realize this and behave accordingly, for now
checking instead:
if (pid && pid2 && pid != nsinfo__nstgid(nsi))
Translating to:
if (3308868 && 1 && 3308868 != 3308868)
Will make the jitdump mmap to be considered valid and processed.
The jitdump is described in:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/perf/Documentation/jitdump-specification.txt
Now we end up with the expected flurry of MMAPs, one per jitted function
transformed into a little ELF file that should then be processable by
the other perf features, like code annotation:
[acme@toolbox a]$ echo $JITDUMPDIR
/tmp/.debug/jit
[acme@toolbox a]$
First use 'perf inject':
⬢ [acme@toolbox a]$ time perf inject -i perf.data -o acme-perf-injected.data -j
Then look at the PERF_RECORD_MMAP events in the result file, that went
thru the JIT map file:
⬢ [acme@toolbox a]$ ls -la /tmp/*.map
-rw-r--r--. 1 acme acme 2989559 Nov 27 16:11 /tmp/perf-3308868.map
[acme@toolbox a]$
It is a symbol table:
⬢ [acme@toolbox a]$ head /tmp/*.map
0x00007fb8bda5c1a0 0x00000000000000d0 java.lang.String java.lang.module.ModuleDescriptor.name()
0x00007fb8bda5c4a0 0x0000000000000178 int java.lang.StringLatin1.hashCode(byte[])
0x00007fb8bda5c9a0 0x00000000000000d0 java.lang.String org.springframework.boot.context.config.ConfigDataLocation.getValue()
0x00007fb8bda5cca0 0x00000000000000d0 java.lang.module.ModuleDescriptor java.lang.module.ModuleReference.descriptor()
0x00007fb8bda5cfa0 0x00000000000000d0 java.lang.Object java.util.KeyValueHolder.getKey()
0x00007fb8bda5d2a0 0x00000000000000d0 java.lang.Object java.util.KeyValueHolder.getValue()
0x00007fb8bda5d5a0 0x0000000000000218 boolean jdk.internal.misc.Unsafe.compareAndSetReference(java.lang.Object, long, java.lang.Object, java.lang.Object)
0x00007fb8bda5d9a0 0x00000000000001f0 boolean jdk.internal.misc.Unsafe.compareAndSetLong(java.lang.Object, long, long, long)
0x00007fb8bda5dda0 0x00000000000001f8 void java.lang.System.arraycopy(java.lang.Object, int, java.lang.Object, int, int)
0x00007fb8bda5e1a0 0x00000000000001e8 int java.lang.Object.hashCode()
⬢ [acme@toolbox a]$
As specified in:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/perf/Documentation/jit-interface.txt
This was collected from inside the container, so came as
/tmp/perf-1.map.
To make perf, running outside the container to use it we need to copy it
to /tmp/perf-3308868.map.
This is another logic that has to be added to perf to work on this
scenario of running outside the container but processing things created
by the hava agent running inside the container.
With all this in place we get to:
⬢ [acme@toolbox a]$ perf report -D -i acme-perf-injected.data | \
grep PERF_RECORD_MMAP > /tmp/acme-perf-injected.data.mmaps ; \
wc -l /tmp/acme-perf-injected.data.mmaps
44182 /tmp/acme-perf-injected.data.mmaps
⬢ [acme@toolbox a]$ tail /tmp/acme-perf-injected.data.mmaps
1030266786574466 0x7bc9e0 [0x98]: PERF_RECORD_MMAP2 1/78: [0x7fb8c0ceb1c0(0x8d0) @ 0x80 00:2c 238715 1]: --xs /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43989.so
1030266795288774 0x7bca78 [0x98]: PERF_RECORD_MMAP2 1/78: [0x7fb8c0cecc00(0x7e8) @ 0x80 00:2c 238716 1]: --xs /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43990.so
1030266895967339 0x7bcb10 [0x98]: PERF_RECORD_MMAP2 1/78: [0x7fb8c0cee500(0x3328) @ 0x80 00:2c 238717 1]: --xs /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43991.so
1030266915748306 0x7bcba8 [0x98]: PERF_RECORD_MMAP2 1/78: [0x7fb8c0aae0a0(0x138) @ 0x80 00:2c 238718 1]: --xs /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43992.so
1030267185851220 0x7bcc40 [0x98]: PERF_RECORD_MMAP2 1/78: [0x7fb8c0cf61e0(0x3b50) @ 0x80 00:2c 238719 1]: --xs /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43993.so
1030267231364524 0x7bccd8 [0x98]: PERF_RECORD_MMAP2 1/78: [0x7fb8c0cfea80(0x14a0) @ 0x80 00:2c 238720 1]: --xs /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43994.so
1030267425498831 0x7bcd70 [0x98]: PERF_RECORD_MMAP2 1/78: [0x7fb8c054b4a0(0x338) @ 0x80 00:2c 238721 1]: --xs /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43995.so
1030267506147888 0x7bce08 [0x98]: PERF_RECORD_MMAP2 1/78: [0x7fb8c0a995c0(0x1e8) @ 0x80 00:2c 238722 1]: --xs /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43996.so
1030268112586116 0x7bcea0 [0x98]: PERF_RECORD_MMAP2 1/78: [0x7fb8c0d02520(0x258) @ 0x80 00:2c 238723 1]: --xs /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43997.so
1030269435398150 0x7bcf38 [0x98]: PERF_RECORD_MMAP2 1/78: [0x7fb8c0d02dc0(0x278) @ 0x80 00:2c 238724 1]: --xs /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43998.so
⬢ [acme@toolbox a]$
And if we look at those tiny ELF files generated by the jitdump code
used by 'perf inject' we see:
⬢ [acme@toolbox a]$ file /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43989.so
/tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43989.so: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV), statically linked, BuildID[sha1]=790591db95a77d644657dfe5058658b200000000, with debug_info, not stripped
⬢ [acme@toolbox a]$ file /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43990.so
/tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43990.so: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV), statically linked, BuildID[sha1]=762f932acbee53a22638bf4c2b86780200000000, with debug_info, not stripped
⬢ [acme@toolbox a]$
⬢ [acme@toolbox a]$ ls -la /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43989.so /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43990.so
-rw-r--r--. 1 acme acme 9432 Nov 29 10:56 /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43989.so
-rw-r--r--. 1 acme acme 7504 Nov 29 10:56 /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43990.so
⬢ [acme@toolbox a]$
And:
⬢ [acme@toolbox a]$ objdump -dS /tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43990.so | head -20
/tmp/.debug/jit/java-jit-20241126.XXTxEIOn/jitted-1-43990.so: file format elf64-x86-64
Disassembly of section .text:
0000000000000080 <Lredacted/REDACTED/REDACTED/logging/RedactedRedacted;Redacted(Lredacted/REDACTED/REDACTED/redactedRedacted/Redacted;)V>:
80: 44 8b 56 08 mov 0x8(%rsi),%r10d
84: 49 c1 e2 03 shl $0x3,%r10
88: 49 3b c2 cmp %r10,%rax
8b: 0f 85 6f 15 83 fc jne fffffffffc831600 <Lredacted/REDACTED/REDACTED/redacted/RedactedRedactedRedacted;Redacted(Lredacted/Redacted/Redacted/redactedRedacted/Redacted;)V+0xfffffffffc831580>
91: 66 66 90 data16 xchg %ax,%ax
94: 0f 1f 84 00 00 00 00 nopl 0x0(%rax,%rax,1)
9b: 00
9c: 66 66 66 90 data16 data16 xchg %ax,%ax
a0: 89 84 24 00 c0 fe ff mov %eax,-0x14000(%rsp)
a7: 55 push %rbp
a8: 48 8b ec mov %rsp,%rbp
ab: 48 83 ec 40 sub $0x40,%rsp
af: 48 89 34 24 mov %rsi,(%rsp)
⬢ [acme@toolbox a]$
The thing now being investigated is why we can't annotate anything here,
maybe that JITDUMPDIR is getting in the way:
⬢ [acme@toolbox a]$ perf annotate --stdio2 -i acme-perf-injected.data 'java.lang.String com.fasterxml.jackson.core.sym.CharsToNameCanonicalizer.findSymbol(char[], int, int, int)'
Error:
Couldn't annotate java.lang.String com.fasterxml.jackson.core.sym.CharsToNameCanonicalizer.findSymbol(char[], int, int, int):
Internal error: Invalid -1 error code
⬢ [acme@toolbox a]$
In the tests I performed while merging this patch:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=6d518ac7be6223811ab947897273b1bbef846180
It works, but then there was no JITDUMPDIR involved:
/home/acme/.debug/jit/java-jit-20241127.XXF1SRgN/jitted-3912413-4191.so
⬢ [acme@toolbox perf-tools-next]$ perf report --call-graph none --no-child -i perf-injected.data | grep jitted- | head
1.36% java jitted-3912413-54.so [.] Interpreter
0.30% C1 CompilerThre jitted-3912413-1.so [.] flush_icache_stub
0.18% java jitted-3912413-4184.so [.] org.apache.fop.fo.properties.PropertyMaker.get(int, org.apache.fop.fo.PropertyList, boolean, boolean)
0.18% java jitted-3912413-4177.so [.] org.apache.fop.layoutmgr.inline.TextLayoutManager.getNextKnuthElements(org.apache.fop.layoutmgr.LayoutContext, int)
0.13% java jitted-3912413-3845.so [.] java.text.DecimalFormat.subformatNumber(java.lang.StringBuffer, java.text.Format$FieldDelegate, boolean, boolean, int, int, int, int)
0.11% java jitted-3912413-4191.so [.] org.apache.fop.fo.FObj.addChildNode(org.apache.fop.fo.FONode)
0.09% java jitted-3912413-2418.so [.] org.apache.fop.fo.XMLWhiteSpaceHandler.handleWhiteSpace()
0.08% Reference Handl jitted-3912413-54.so [.] Interpreter
0.08% java jitted-3912413-3326.so [.] org.apache.xmlgraphics.fonts.Glyphs.stringToGlyph(java.lang.String)
0.08% java jitted-3912413-3953.so [.] org.apache.fop.layoutmgr.BreakingAlgorithm.considerLegalBreak(org.apache.fop.layoutmgr.KnuthElement, int)
⬢ [acme@toolbox perf-tools-next]$
And then:
⬢ [acme@toolbox perf-tools-next]$ perf annotate --stdio2 -i perf-injected.data 'org.apache.fop.layoutmgr.inline.TextLayoutManager.getNextKnuthElements(org.apache.fop.layoutmgr.LayoutContext, int)' | head -20
Samples: 8 of event 'cpu_atom/cycles/Pu', 4000 Hz, Event count (approx.): 8112794, [percent: local period]
org.apache.fop.layoutmgr.inline.TextLayoutManager.getNextKnuthElements(org.apache.fop.layoutmgr.LayoutContext, int)() /home/acme/.debug/jit/java-jit-20241127.XXF1SRgN/jitted-3912413-4177.so
Percent 0x80 <org.apache.fop.layoutmgr.inline.TextLayoutManager.getNextKnuthElements(org.apache.fop.layoutmgr.LayoutContext, int)>:
nop
movl 0x8(%rsi),%r10d
cmpl 0x8(%rax),%r10d
→ jne 0
movl %eax,-0x14000(%rsp)
pushq %rbp
subq $0xb0,%rsp
nop
cmpl $0x3,0x20(%r15)
↓ jne 7037
2e: movl %ecx,0x28(%rsp)
movq %rdx,%rbp
movl 0x64(%rdx),%ebx
cmpb $0x0,0x38(%r15)
↓ jne 3a44
movq %rsi,0x30(%rsp)
48: movq 0x30(%rsp),%r10
⬢ [acme@toolbox perf-tools-next]$
No source code nor line numbers, that I saw in another build of perf for
RHEL9, for the same workload described in the cset above (a publicly
available java benchmark), so something to investigate on perf upstream
running on fedora, maybe some quirk with the jdk used when building perf
for RHEL 9 and for Fedora 40.
A related patch that should have make this all work is:
"perf inject jit: Add namespaces support"
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=67dec926931448d688efb5fe34f7b5a22470fc0a
But we still need to polish this some more, maybe there are differences
in the agent used in NodeJS with --perf-prof and the jvmti one we're
using.
Hopefully describing all the steps while we investigate this case will
help us improve perf support for profiling JITed environments running in
containers while profiling from inside and outside it.
Reported-by: Francesco Nigro <fnigro@redhat.com>
Reported-by: Ilan Green <igreen@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Clark Williams <williams@redhat.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Yonatan Goldschmidt <yonatan.goldschmidt@granulate.io>
Link: https://lore.kernel.org/r/20241206204828.507527-3-acme@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Depending on how vmlinux.lds is written, _etext might be the very first
data symbol instead of the very last text symbol.
Don't require it to be a text symbol, accept any symbol type.
Comitter notes:
See the first Link for further discussion, but it all boils down to
this:
---
# grep -e _stext -e _etext -e _edata /proc/kallsyms
c0000000 T _stext
c08b8000 D _etext
So there is no _edata and _etext is not text
$ ppc-linux-objdump -x vmlinux | grep -e _stext -e _etext -e _edata
c0000000 g .head.text 00000000 _stext
c08b8000 g .rodata 00000000 _etext
c1378000 g .sbss 00000000 _edata
---
Fixes: ed9adb2035b5be58 ("perf machine: Read also the end of the kernel")
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.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: linuxppc-dev@lists.ozlabs.org
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/r/b3ee1994d95257cb7f2de037c5030ba7d1bed404.1736327613.git.christophe.leroy@csgroup.eu
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Since commit 659ad3492b913c90 ("perf maps: Switch from rbtree to lazily
sorted array for addresses"), perf doesn't display anymore kernel
symbols on powerpc, allthough it still detects them as kernel addresses.
# Overhead Command Shared Object Symbol
# ........ .......... ............. ......................................
#
80.49% Coeur main [unknown] [k] 0xc005f0f8
3.91% Coeur main gau [.] engine_loop.constprop.0.isra.0
1.72% Coeur main [unknown] [k] 0xc005f11c
1.09% Coeur main [unknown] [k] 0xc01f82c8
0.44% Coeur main libc.so.6 [.] epoll_wait
0.38% Coeur main [unknown] [k] 0xc0011718
0.36% Coeur main [unknown] [k] 0xc01f45c0
This is because function maps__find_next_entry() now returns current
entry instead of next entry, leading to kernel map end address getting
mis-configured with its own start address instead of the start address
of the following map.
Fix it by really taking the next entry, also make sure that entry
follows current one by making sure entries are sorted.
Fixes: 659ad3492b913c90 ("perf maps: Switch from rbtree to lazily sorted array for addresses")
Reviewed-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
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: 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>
Link: https://lore.kernel.org/r/2ea4501209d5363bac71a6757fe91c0747558a42.1736329923.git.christophe.leroy@csgroup.eu
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Sometimes users also want to see average latency as well as histogram.
Display latency statistics like avg, max, min at the end.
$ sudo ./perf ftrace latency -ab -T synchronize_rcu -- ...
# DURATION | COUNT | GRAPH |
0 - 1 us | 0 | |
1 - 2 us | 0 | |
2 - 4 us | 0 | |
4 - 8 us | 0 | |
8 - 16 us | 0 | |
16 - 32 us | 0 | |
32 - 64 us | 0 | |
64 - 128 us | 0 | |
128 - 256 us | 0 | |
256 - 512 us | 0 | |
512 - 1024 us | 0 | |
1 - 2 ms | 0 | |
2 - 4 ms | 0 | |
4 - 8 ms | 0 | |
8 - 16 ms | 1 | ##### |
16 - 32 ms | 7 | ######################################## |
32 - 64 ms | 0 | |
64 - 128 ms | 0 | |
128 - 256 ms | 0 | |
256 - 512 ms | 0 | |
512 - 1024 ms | 0 | |
1 - ... s | 0 | |
# statistics (in usec)
total time: 171832
avg time: 21479
max time: 30906
min time: 15869
count: 8
Committer testing:
root@number:~# perf ftrace latency -nab --bucket-range 100 --max-latency 512 -T switch_mm_irqs_off sleep 1
# DURATION | COUNT | GRAPH |
0 - 100 ns | 314 | ## |
100 - 200 ns | 1843 | ############# |
200 - 300 ns | 1390 | ########## |
300 - 400 ns | 844 | ###### |
400 - 500 ns | 480 | ### |
500 - 512 ns | 315 | ## |
512 - ... ns | 16 | |
# statistics (in nsec)
total time: 2448936
avg time: 387
max time: 3285
min time: 82
count: 6328
root@number:~#
Reviewed-by: James Clark <james.clark@linaro.org>
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: 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>
Link: https://lore.kernel.org/r/20250107224352.1128669-1-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
The existing EBUSY strerror message is:
The sys_perf_event_open() syscall returned with 16 (Device or resource busy) for event (intel_bts//).
"dmesg | grep -i perf" may provide additional information.
The dmesg won't be useful. What is more useful is knowing what
processes are potentially using the PMU, which some procfs scanning can
reveal. When parallel testing tests/shell/stat_all_pmu.sh this yields:
Testing intel_bts//
Error:
The PMU intel_bts counters are busy and in use by another process.
Possible processes:
2585882 perf list
2585902 perf list -j -o /tmp/__perf_test.list_output.json.KF9MY
2585904 perf list
2585911 perf record -e task-clock --filter period > 1 -o /dev/null --quiet true
2585912 perf list
2585915 perf list
2586042 /tmp/perf/perf record -asdg -e cpu-clock -o /tmp/perftool-testsuite_report.dIF/perf_report/perf.data -- sleep 2
2589078 perf record -g -e task-clock:u -o - perf test -w noploop
2589148 /tmp/perf/perf record --control=fifo:control,ack -e cpu-clock -m 1 sleep 10
2589379 perf --buildid-dir /tmp/perf.debug.Umx record --buildid-all -o /tmp/perf.data.YBm /tmp/perf.ex.MD5.ZQW
2589568 perf record -o /tmp/__perf_test.program.mtcZH/perf.data --branch-filter any,save_type,u -- perf test -w brstack
2589649 perf record --per-thread -o /tmp/__perf_test.perf.data.5d3dc perf test -w thloop
2589898 perf record -o /tmp/perf-test-script.BX2b27Dcnj/pp-perf.data --sample-cpu uname
Which gets a little closer to finding the issue.
Committer testing:
root@number:~#
root@number:~# grep -m1 "model name" /proc/cpuinfo
model name : Intel(R) Core(TM) i7-14700K
root@number:~#
Before:
root@number:~# perf stat -e intel_bts// &
[1] 197954
root@number:~# perf test "perf all PMU test"
124: perf all PMU test : FAILED!
root@number:~# perf test -v "perf all PMU test" |& tail
Testing i915/vecs0-busy/
Testing i915/vecs0-sema/
Testing i915/vecs0-wait/
Testing intel_bts//
Unexpected signal in main
Error:
The sys_perf_event_open() syscall returned with 16 (Device or resource busy) for event (intel_bts//).
"dmesg | grep -i perf" may provide additional information.
---- end(-1) ----
124: perf all PMU test : FAILED!
root@number:~#
After:
root@number:~# perf stat -e intel_bts// &
[1] 200195
root@number:~# perf test "perf all PMU test"
123: perf all PMU test : FAILED!
root@number:~# perf test -v "perf all PMU test" |& tail
Testing i915/vecs0-wait/
Testing intel_bts//
Unexpected signal in main
Error:
The PMU intel_bts counters are busy and in use by another process.
Possible processes:
200195 perf stat -e intel_bts//
2319766 /root/bin/perf top --stdio
---- end(-1) ----
123: perf all PMU test : FAILED!
root@number:~#
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: Chun-Tse Shao <ctshao@google.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>
Cc: Weilin Wang <weilin.wang@intel.com>
Cc: Ze Gao <zegao2021@gmail.com>
Change-Id: Ie1ed8688286c44e8f44a35e98fed8be3e2a344df
Link: https://lore.kernel.org/r/20241106003007.2112584-1-ctshao@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Not all of these are "state" so separate them into two sections. Rename
and document to make all clearer.
Signed-off-by: James Clark <james.clark@linaro.org>
Tested-by: Ian Rogers <irogers@google.com>
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: 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: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Yicong Yang <yangyicong@hisilicon.com>
Link: https://lore.kernel.org/r/20241112160048.951213-6-james.clark@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Rename 'prefix' to 'timestamp' because that's all it does, except in
iostat mode where it's slightly overloaded, but still includes a
timestamp. This reveals a problem with iostat and JSON mode so document
this.
Make it more explicit that these are printed in interval mode by
changing 'if (prefix)' to 'if (interval)' which reveals an unnecessary
'else if (... && !interval)' which can be removed.
Signed-off-by: James Clark <james.clark@linaro.org>
Tested-by: Ian Rogers <irogers@google.com>
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: 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: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Yicong Yang <yangyicong@hisilicon.com>
Link: https://lore.kernel.org/r/20241112160048.951213-5-james.clark@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Despite the name new_line_metric doesn't make a new line, it actually
does nothing. Change it to NULL to avoid confusion.
Signed-off-by: James Clark <james.clark@linaro.org>
Tested-by: Ian Rogers <irogers@google.com>
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: 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: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Yicong Yang <yangyicong@hisilicon.com>
Link: https://lore.kernel.org/r/20241112160048.951213-4-james.clark@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
We decided to hide NULL metric-units rather than showing it as "(null)"
when a dependent event for a metric doesn't exist. But on hybrid systems
if the process doesn't hit a PMU you get an empty string metric unit
instead. To make it consistent change all empty strings to NULL.
Note that metric-threshold is already hidden in this case without this
change.
Where a process only runs on cpu_core and never hits cpu_atom:
Before:
$ perf stat -j -- true
...
{"counter-value" : "<not counted>", "unit" : "", "event" : "cpu_atom/branch-misses/", "event-runtime" : 0, "pcnt-running" : 0.00, "metric-value" : "0.000000", "metric-unit" : ""}
{"counter-value" : "6326.000000", "unit" : "", "event" : "cpu_core/branch-misses/", "event-runtime" : 293786, "pcnt-running" : 100.00, "metric-value" : "3.553394", "metric-unit" : "of all branches", "metric-threshold" : "good"}
...
After:
...
{"counter-value" : "<not counted>", "unit" : "", "event" : "cpu_atom/branch-misses/", "event-runtime" : 0, "pcnt-running" : 0.00}
{"counter-value" : "5778.000000", "unit" : "", "event" : "cpu_core/branch-misses/", "event-runtime" : 282240, "pcnt-running" : 100.00, "metric-value" : "3.226797", "metric-unit" : "of all branches", "metric-threshold" : "good"}
...
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: James Clark <james.clark@linaro.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Ian Rogers <irogers@google.com>
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: 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: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Yicong Yang <yangyicong@hisilicon.com>
Link: https://lore.kernel.org/r/20241112160048.951213-3-james.clark@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Now that printing metric-value and metric-unit is optional,
print_running_json() shouldn't add the comma in case it becomes
trailing.
Replace all manual JSON comma stuff with a json_out() function that uses
the existing os->first tracking and auto inserts a comma if it's needed.
Update the test to handle that two of the fields can be missing.
This fixes the following test failure on Cortex A57 where the branch
misses metric is missing a required event:
$ perf test -vvv "json output"
106: perf stat JSON output linter:
--- start ---
test child forked, pid 665682
Checking json output: no args Test failed for input:
{"counter-value" : "3112.000000", "unit" : "",
"event" : "armv8_pmuv3_1/branch-misses/",
"event-runtime" : 20699340, "pcnt-running" : 100.00, }
...
json.decoder.JSONDecodeError: Expecting property name enclosed in
double quotes: line 12 column 144 (char 2109)
---- end(-1) ----
106: perf stat JSON output linter : FAILED!
Fixes: e1cc918b6cfd1206 ("perf stat: Drop metric-unit if unit is NULL")
Signed-off-by: James Clark <james.clark@linaro.org>
Tested-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.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: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Yicong Yang <yangyicong@hisilicon.com>
Link: https://lore.kernel.org/r/20241112160048.951213-2-james.clark@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
is_executable_file() has been unused since 2022's commit
7391db6459388d47 ("perf test: Refactor shell tests allowing subdirs")
Remove it.
Signed-off-by: Dr. David Alan Gilbert <linux@treblig.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Carsten Haitzler <carsten.haitzler@arm.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: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20241222215831.283248-1-linux@treblig.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
