| Age | Commit message (Collapse) | Author |
|---|
|
The updated events were published in:
https://github.com/intel/perfmon/commit/13983cd535d18b2bfd86a3b9daa374039f78a836
Signed-off-by: Chun-Tse Shao <ctshao@google.com>
Reviewed-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The updated events were published in:
https://github.com/intel/perfmon/commit/b23ebe7bc25add0c835565e4bc87e063cb620a02
Signed-off-by: Chun-Tse Shao <ctshao@google.com>
Reviewed-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The updated events were published in:
https://github.com/intel/perfmon/commit/7a14cc8feaf86772deb6708e96c8e9fee6d5b1ca
Signed-off-by: Chun-Tse Shao <ctshao@google.com>
Reviewed-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The updated events were published in:
https://github.com/intel/perfmon/commit/7a14cc8feaf86772deb6708e96c8e9fee6d5b1ca
Signed-off-by: Chun-Tse Shao <ctshao@google.com>
Reviewed-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
In evsel__tp_format(), when trace_event__tp_format*() returns an error,
IS_ERR() checks the local variable 'tp_format', but PTR_ERR() incorrectly
uses 'evsel->tp_format' which hasn't been assigned yet.
Fix this by using PTR_ERR(tp_format) to extract the error code from the
correct variable.
Fixes: 6c8310e8380d ("perf evsel: Allow evsel__newtp without libtraceevent")
Signed-off-by: Hongling Zeng <zenghongling@kylinos.cn>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Use MAP_FAILED instead of NULL to detect mmap errors, and fix the
slots_p variable name typo in the sample code.
Signed-off-by: Hongfu Li <lihongfu@kylinos.cn>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
New OMR MSRs are introduced for OMR events on DMR and NVL. Perf continues
to reuse the existing offcore_rsp attribute to encode the MSR value,
similar to existing OCR event handling.
Add the corresponding OMR MSR mappings in lookup_msr() so jevents can
translate these events and generate the correct offcore_rsp attribute.
Link: https://lore.kernel.org/all/20260114011750.350569-2-dapeng1.mi@linux.intel.com/
Signed-off-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Remove duplicate inclusion of util/term.h in builtin-test.c to clean up
redundant code.
Signed-off-by: Chen Ni <nichen@iscas.ac.cn>
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
With Zstd compression enabled ('perf record -z'), a single mmap push
whose compressed output exceeds the maximum record size makes
zstd_compress_stream_to_records() emit several PERF_RECORD_COMPRESSED2
records back to back. record__pushfn() however rewrote only the first
record's header to describe the whole blob as one record:
event->data_size = compressed - sizeof(struct perf_record_compressed2);
event->header.size = PERF_ALIGN(compressed, sizeof(u64));
padding = event->header.size - compressed;
...
record__write(rec, map, &pad, padding);
perf_event_header::size is a __u16, so once the compressed blob no
longer fits in it the header.size assignment truncates and 'padding'
(size_t) underflows. write() is then handed that bogus length and fails
with EFAULT, aborting the recording:
failed to write perf data, error: Bad address
The bytes that did reach the file are mis-framed, so reading it back
cannot be decompressed.
This is easy to hit with a high event rate and a large buffer, e.g.:
perf record -z -F max -m 32M --per-thread -- perf test -w thloop 5 1
The single-record fixup is wrong by construction: because header.size is
16 bits a compressed record cannot exceed 64KB, so the compressor must
split a push into a chain of records, and the session reader already
consumes them as such.
Frame each record where it is produced instead: make
process_comp_header() set the per-record data_size, 8-byte-align
header.size and zero the trailing padding, and let record__pushfn()
write the resulting blob, as the AIO path already does. Reduce
max_record_size by sizeof(u64) so the per-record alignment padding
cannot push header.size past its u16 field. process_comp_header()
returns -1 when that padding would not fit the space left in 'dst', so
the compressor stops instead of overrunning the output buffer.
There is no on-disk format change; a perf.data written by the fixed tool
is still read by existing perf.
Fixes: 208c0e168344 ("perf record: Add 8-byte aligned event type PERF_RECORD_COMPRESSED2")
Reported-by: Farid Zakaria <fmzakari@meta.com>
Signed-off-by: Dmitry Ilvokhin <d@ilvokhin.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
process_comp_header() is called from zstd_compress_stream_to_records()
twice per record: once with data_size == 0 to write the record header,
and once with the payload size to finalize it. It returns the increment
it was passed, and the loop separately decides whether a record still
fits by comparing the remaining 'dst_size' against the header size.
With the fit check split from the code that writes the record,
process_comp_header() cannot reject a record on its own, so any bytes it
writes into 'dst' have to be bounds-checked by the caller instead of
where they are produced.
Pass the space left in 'dst' to process_comp_header(), let it return the
number of bytes written or -1 when the header does not fit, and account
the compressed payload in the loop.
No functional change intended.
Signed-off-by: Dmitry Ilvokhin <d@ilvokhin.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The recent commit caused a failure in make build-test for static builds.
Let's not pass -static the option to dlfilters which is dynamically
loaded as it's hard-coded with -shared even for static builds.
Tested-by: Leo Yan <leo.yan@arm.com>
Cc: Trevor Allison <tallison@redhat.com>
Fixes: e1065ed188cf ("perf build: Add LDFLAGS to dlfilters .so link")
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
metricgroup__copy_metric_events() allocates a new metric expression and
duplicates metric_name before linking the expression into the destination
metric event.
Free new_expr when strdup() fails, and free the duplicated metric_name on
the later error paths.
Fixes: b85a4d61d302 ("perf metric: Allow modifiers on metrics")
Signed-off-by: Yu Peng <pengyu@kylinos.cn>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
This commit updates the .clang-format rules for the perf tools.
It prioritizes Python.h sorting and expands the regex lists for C
standard library and OS specific headers.
- Prioritize Python.h by sorting it first (Priority -1), ensuring it
is sorted before even the main header (Priority 0).
- Keep other header priorities (1 and 2) to maintain their relative
order.
- Sanity check and complete "C Standard Library Headers" up to C23
(added stdbit.h, stdckdint.h).
- Sanity check and complete "OS/System-Specific Headers" based on
actual usage in tools/perf (added asm-generic, uapi, and various
flat POSIX/system headers).
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Under system-wide (-a) parallel streaming mode (--threads=cpu),
background recording threads can be inundated by a continuous
firehose of hardware samples generated by the OS. In this state,
a background thread's local hit count remains unequal to its
sample count, causing it to bypass the blocking fdarray__poll()
call entirely on each iteration of its recording loop.
Because the termination check relies on the POLLHUP event status
populated specifically by fdarray__poll(), bypassing it prevents
the background thread from ever recognizing that its control pipe
was closed by the main thread. This traps the background thread
in an infinite recording loop, hanging the main thread indefinitely
as it awaits a termination acknowledgment that never arrives.
Ensure teardown completion by adding explicit evlist__disable()
calls in the main thread's cleanup paths at out_child: and
out_child_no_flush:. Additionally, patch fdarray__filter() to
respect the fdarray_flag__nonfilterable flag, preventing it
from incorrectly setting the background thread's control pipe
file descriptor to -1 and clearing its revents mask upon
processing termination POLLHUP signals.
Fixes: f94563fac269 ("perf record: fix poll storm when monitored threads exit")
Assisted-by: Antigravity:gemini-3.5-flash
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
In print_capstone_detail(), maps__find() is used to locate the kernel
map. This function increments the reference count of the found map
object. However, the current implementation fails to call map__put()
after the map is no longer needed, leading to a reference count leak.
Fix this by adding a map__put(map) call to properly release the
reference after use.
Fixes: 92dfc59463d5 ("perf annotate: Add symbol name when using capstone")
Signed-off-by: Tengda Wu <wutengda@huaweicloud.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Commit ae42a2a2a3ae ("perf tests: Speed up lock contention analysis shell test")
in linux-next heavily optimized the test runtimes by switching the workload from
the default of 10 process groups down to 1 (`perf bench sched messaging -g 1`).
However, this change inadvertently dropped the original `-p` flag, causing the
benchmark to default to `socketpair()` instead of `pipe()`. While `socketpair()`
still generates some lock events on x86, it fails to trigger enough samples on
architectures like s390, causing the test suite to fail due to lack of captured
data.
Restore the omitted `-p` pipe flag. The test retains the massive speedups
achieved through the `-g 1` scaling, while producing a massive density of lock
events across all architectures to fully satisfy the BPF trace filtering logic.
Fixes: ae42a2a2a3ae ("perf tests: Speed up lock contention analysis shell test")
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Richter <tmricht@linux.ibm.com>
Assisted-by: Antigravity:gemini-3.1-pro
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The current kallsyms detection in dso__is_kallsyms() uses the
dso_binary_type enum which fixes the issue of kallsyms being cached in
the build-id cache for out-of-tree modules.
However, during build-id injection in perf record/inject, dso_binary_type
has not been explicitly set yet,so dso__binary_type() returns
DSO_BINARY_TYPE__NOT_FOUND instead of DSO_BINARY_TYPE__KALLSYMS for the
kernel DSO. The current check then fails to identify it as kallsyms,
causing build-id symlinks to not be created in ~/.debug/.build-id/ and
perf archive to fail with "Cannot stat" errors.
Steps to reproduce the issue:
1. rm -rf ~/.debug/.build-id
2. perf record sleep 1
3. perf archive
Fix by falling back to matching long_name against the known kallsyms
strings explicitly when binary_type is not yet set
(== DSO_BINARY_TYPE__NOT_FOUND). Use strcmp() for exact matching of
fixed names and strict validation for guest kallsyms with embedded PID
to prevent path traversal attacks.
Fixes: ebf0b332732d ("perf dso: fix dso__is_kallsyms() check")
Signed-off-by: Tanushree Shah <tshah@linux.ibm.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Specifying --field-separator option is stating you want CSV output.
Passing both --field-separator and --json-output is then stating
you want output to be in CSV and JSON format at same time.
Currently this combination is not rejected, and the outcome
is a malformed combination of CSV and JSON output.
This is because of inconsistencies in various printing functions,
some of them have if-else chains that start with
"Should I print JSON?", and some start with "Should I print CSV?".
Example of current output:
$ tools/perf/perf stat -x , -j -e cpu-migrations true
{"counter-value" : "0.000000", "unit" : "", "event" : "cpu-migrations", "event-runtime" : 474817, "pcnt-running" : 100.00,,
Instead reject the option combination,
with a helpful error message and non-zero exit code.
Example of new output:
$ tools/perf/perf stat -x , -j true
cannot use both --field-separator and --json-output
Usage: perf stat [<options>] [<command>]
-x, --field-separator <separator>
print counts with custom separator
-j, --json-output print counts in JSON format
Signed-off-by: Ivan Lazaric <ivan.lazaric1@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Maintain backward compatibility by reintroducing the events that were
previously removed by commit 047979af3bf6 ("perf vendor events amd:
Update Zen 5 core events"). Also set the deprecated flag and update
the descriptions to point users to the correct alternative.
Reported-by: Ian Rogers <irogers@google.com>
Closes: https://lore.kernel.org/all/CAP-5=fV_czvd-z4N7K+_SabxuOm9UUHRyBxNuchrtAgJL3OqOw@mail.gmail.com/
Fixes: 047979af3bf6 ("perf vendor events amd: Update Zen 5 core events")
Signed-off-by: Sandipan Das <sandipan.das@amd.com>
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Unlike the in-kernel trace_seq which uses a statically allocated buffer,
the userspace traceevent library's trace_seq uses a dynamically allocated
one. Therefore, every trace_seq_init() call must be paired with a
trace_seq_destroy(), otherwise it produces a memory leak.
In process_sample_event(), a trace_seq is initialized for each field when
formatting tracepoint raw_data, but the matching trace_seq_destroy() is
never called, leaking memory for every field of every sample processed.
Add the missing trace_seq_destroy() after using the trace_seq buffer to
properly free the allocated memory.
Detected with Valgrind on a perf.data file with 2,729 tracepoint samples:
Before: definitely lost: 55,537,664 bytes in 13,559 blocks
After: definitely lost: 0 bytes in 0 blocks
Fixes: 9d895e468429 ("perf data: Add tracepoint fields when converting to JSON")
Signed-off-by: Tanushree Shah <tshah@linux.ibm.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
When `perf record` samples a multi-threaded process and one of the
target threads exits during the session, perf itself may start burning
100% CPU (up to 200% across two cores) until the session ends. A
single dead fd is sufficient to trigger this; it can be reproduced with
15 pthreads in a compute loop where one thread exits halfway through.
The root cause is two independent instances of the same defect: dead
perf_event ring-buffer fds are left in a pollfd array. When a monitored
thread exits, the kernel closes its ring-buffer fd, which then returns
POLLHUP. POSIX specifies that poll() always reports POLLHUP and POLLERR
regardless of the events mask, so any dead fd left in the array makes
poll() return immediately every time, spinning in a tight loop:
3 seconds: 256,600 poll() calls, 0 context switches, only 21 write()
Woken up count goes from ~0 to 1,300,000+
There are two affected poll paths, fixed together here:
1. Record main loop, via fdarray__filter() (tools/lib/api/fd/array.c).
Since commit 59b4412f27f1 ("libperf: Avoid internal moving of
fdarray fds") it only zeroes events/revents without setting fd to
-1, so poll() keeps reporting POLLHUP for the entry. Setting
fd = -1 makes poll() skip it, matching the pattern already used in
the control-fd path at tools/perf/builtin-record.c:1673.
2. BPF sideband thread, perf_evlist__poll_thread()
(tools/perf/util/sideband_evlist.c). This thread polls for
PERF_RECORD_BPF_EVENT but, unlike the main record loop, never calls
fdarray__filter() at all, so dead fds accumulate forever and it
spins at 100% CPU:
Before fix: dJiffies=101, wchan=0 (running)
After fix: dJiffies=0, wchan=do_sys_poll (blocking)
Fixed by calling the existing evlist__filter_pollfd() helper after
evlist__poll(), mirroring the main record loop. <poll.h> is
included for the POLLERR/POLLHUP macros (previously unused there).
The two fixes compose: fix 1 makes poll() ignore dead fds (fd=-1); fix
2 ensures the sideband thread actually performs the filtering. Both
paths are affected in all kernels from v5.1/v5.9 to the current master
(7.2-rc1); the source of both functions is byte-identical across them.
BPF event recording is preserved: after the fix, perf.data still
contains PERF_RECORD_BPF_EVENT records and bpf_prog_info entries.
Verified on perf 6.1.76, 6.6.143 and 7.2-rc1 with a minimal reproducer
(Woken up 1,300,000 -> 3, CPU 100% -> 0%) and an A/B orthogonal test:
keeping the unpatched binary but preventing the target thread from
exiting also makes the storm disappear, confirming the trigger.
Fixes: 59b4412f27f1 ("libperf: Avoid internal moving of fdarray fds")
Fixes: 657ee5531903 ("perf evlist: Introduce side band thread")
Signed-off-by: Jiawei Sun <abyssmystery@gmail.com>
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Add auto counter reload sampling test to verify that the intended event
records can be captured and the self-reloaded events won't generate any
records.
Signed-off-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Update definitions for the following events.
* PMCx00A - Fix descriptions
* PMCx00C - Add missing unit masks
* PMCx00D - Add missing unit masks and fix descriptions
* PMCx013 - Fix incorrect unit masks
Fixes: 2f42fb0661d9 ("perf vendor events amd: Add Zen 6 core events")
Signed-off-by: Sandipan Das <sandipan.das@amd.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Update definitions for the following events.
* PMCx00A - Add missing unit masks
* PMCx00B - Add missing unit masks and fix descriptions
* PMCx00C - Add missing unit masks
* PMCx00D - Add missing unit masks
* PMCx025 - Add missing unit masks and fix descriptions
Fixes: 45c072f2537a ("perf vendor events amd: Add Zen 5 core events")
Signed-off-by: Sandipan Das <sandipan.das@amd.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Making [parts of] the heap executable is dangerous and is blocked by
SELinux on Fedora/RHEL even for an unconfined user. Replace the malloc()
+ mprotect() combo with just mmap(), creating a private anonymous rwx
mapping, which only requires the more commonly allowed "execmem"
permission under SELinux (things like JIT or regex compilation need it
as well). mmap() with MAP_ANONYMOUS will give us a zeroed mapping that
begins on a page boundary, so the result is equivalent to the original
code even without a memset() or the page-alignment dance.
Verified that the test still passes on a machine with an AMD CPU that
has the "ibs" CPU flag.
Fixes: 35db59fa8ea2 ("perf test amd ibs: Add sample period unit test")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Reviewed-by: Ravi Bangoria <ravi.bangoria@amd.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Add a CoreSight shell test for synthesized callchains.
The test uses the new callchain workload to generate trace and decodes
it with synthesis callchain. It then verifies that the instruction
samples show the expected callchain push and pop.
Use control FIFOs so tracing starts only around the workload, which
keeps the trace data small. The test is limited to with the cs_etm
event available and root permission.
After:
perf test 138 -vvv
138: CoreSight synthesized callchain:
---- start ----
test child forked, pid 35581
Callchain flow matched:
l1=4642868 l2=4642880 l3=4642895 l4=4642919 l5=4670494 l6=4670500 l7=4670520
---- end(0) ----
138: CoreSight synthesized callchain : Ok
Assisted-by: Codex:GPT-5.5
Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Leo Yan <leo.yan@arm.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
CS ETM already records branches into the thread stack, but instruction
samples do not carry synthesized callchains. It misses to support the
callchain and no output with the itrace option 'g'.
Allocate a callchain buffer per queue and use thread_stack__sample()
when synthesizing instruction samples.
Advertise PERF_SAMPLE_CALLCHAIN on the synthetic instruction event.
Allocate one extra callchain entry than requested, as the first entry
is reserved for storing context information.
cs_etm__context() is introduced for handling context packet and update
the thread info and start kernel address for frontend decoding.
After:
perf script --itrace=g16l64i1i
callchain_test 6543 [002] 1 instructions:
ffff800080010c14 vectors+0x414 ([kernel.kallsyms])
aaaad6b60784 do_svc+0x1c (/home/kernel/leoy/test_cs_callchain/callchain_test)
aaaad6b60798 print+0xc (/home/kernel/leoy/test_cs_callchain/callchain_test)
aaaad6b607b0 foo+0xc (/home/kernel/leoy/test_cs_callchain/callchain_test)
aaaad6b607c8 main+0xc (/home/kernel/leoy/test_cs_callchain/callchain_test)
ffff9325225c __libc_start_call_main+0x7c (/usr/lib/aarch64-linux-gnu/libc.so.6)
ffff9325233c call_init+0x9c (inlined)
ffff9325233c __libc_start_main_impl+0x9c (inlined)
aaaad6b60670 _start+0x30 (/home/kernel/leoy/test_cs_callchain/callchain_test)
ffff800080012290 ret_to_user+0x120 ([kernel.kallsyms])
Signed-off-by: Leo Yan <leo.yan@linaro.org>
Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Leo Yan <leo.yan@arm.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The perf script callindent is derived from call stack in thread context,
CS ETM ignores the requirement for callindent without pushing and poping
call stack.
Enable thread-stack when either itrace thread-stack support or last branch
entries are requested, allocate the branch stack storage accordingly, and
feed taken branches to thread_stack__event() whenever thread-stack state
is needed.
When callindent is requested, pass callstack=true to thread_stack__event()
so the common thread-stack code maintains call depth for branch samples.
Before:
perf script -F +callindent
callchain_test 6543 [002] 1 branches: main ffff93252258 __libc_start_call_main+0x78 (/usr/lib/aarch64-linux-gnu/libc.so.6)
callchain_test 6543 [002] 1 branches: foo aaaad6b607c4 main+0x8 (/home/kernel/leoy/test_cs_callchain/callchain_test)
callchain_test 6543 [002] 1 branches: print aaaad6b607ac foo+0x8 (/home/kernel/leoy/test_cs_callchain/callchain_test)
callchain_test 6543 [002] 1 branches: do_svc aaaad6b60794 print+0x8 (/home/kernel/leoy/test_cs_callchain/callchain_test)
callchain_test 6543 [002] 1 branches: vectors aaaad6b60780 do_svc+0x18 (/home/kernel/leoy/test_cs_callchain/callchain_test)
callchain_test 6543 [002] 1 branches: el0t_64_sync_handler ffff80008001159c el0t_64_sync+0x194 ([kernel.kallsyms])
callchain_test 6543 [002] 1 branches: el0_svc ffff800081829194 el0t_64_sync_handler+0x9c ([kernel.kallsyms])
callchain_test 6543 [002] 1 branches: lockdep_hardirqs_off ffff800081828794 el0_svc+0x24 ([kernel.kallsyms])
callchain_test 6543 [002] 1 branches: __this_cpu_preempt_check ffff80008182b348 lockdep_hardirqs_off+0xf0 ([kernel.kallsyms])
After:
callchain_test 6543 [002] 1 branches: main ffff93252258 __libc_start_call_main+0x78 (/usr/lib/aarch64-linux-gnu/libc.so.6)
callchain_test 6543 [002] 1 branches: foo aaaad6b607c4 main+0x8 (/home/kernel/leoy/test_cs_callchain/callchain_test)
callchain_test 6543 [002] 1 branches: print aaaad6b607ac foo+0x8 (/home/kernel/leoy/test_cs_callchain/callchain_test)
callchain_test 6543 [002] 1 branches: do_svc aaaad6b60794 print+0x8 (/home/kernel/leoy/test_cs_callchain/callchain_test)
callchain_test 6543 [002] 1 branches: vectors aaaad6b60780 do_svc+0x18 (/home/kernel/leoy/test_cs_callchain/callchain_test)
callchain_test 6543 [002] 1 branches: el0t_64_sync_handler ffff80008001159c el0t_64_sync+0x194 ([kernel.kallsyms])
callchain_test 6543 [002] 1 branches: el0_svc ffff800081829194 el0t_64_sync_handler+0x9c ([kernel.kallsyms])
callchain_test 6543 [002] 1 branches: lockdep_hardirqs_off ffff800081828794 el0_svc+0x24 ([kernel.kallsyms])
callchain_test 6543 [002] 1 branches: __this_cpu_preempt_check ffff80008182b348 lockdep_hardirqs_off+0xf0 ([kernel.kallsyms])
Signed-off-by: Leo Yan <leo.yan@linaro.org>
Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Leo Yan <leo.yan@arm.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Perf resets the CoreSight decoder when moving to a new AUX trace buffer,
this causes trace discontinunity globally.
For callchain synthesis, keeping thread-stack state after decoder reset
can leave stale call/return history attached to threads that are decoded
later, producing incorrect synthesized callchains.
Flush all host thread stacks after a decoder reset. When virtualization
is present, flush the guest thread stacks as well.
Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Leo Yan <leo.yan@arm.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
CS ETM maintains its own circular array for last branch entries, with
local helpers to update, copy and reset the branch stack. This
duplicates logic already provided by the common code.
Record taken branches with thread_stack__event() and synthesize
PERF_SAMPLE_BRANCH_STACK data with thread_stack__br_sample(). This
removes the private last_branch_rb buffer and its position tracking.
This also makes the branch history state belong to the thread rather
than the trace queue. That is a better fit for CoreSight traces where
a trace queue can effectively be CPU scoped, while call/return history
is per thread.
Keep the buffer number updated via thread_stack__set_trace_nr(), which
is used when exporting samples to Python scripts. Pass callstack=false
for now; synthesized callchains are added by a later patch.
The output should remain same, except that be->flags.predicted is no
longer set. Since CoreSight trace does not provide branch prediction
information, clearing the flag avoids confusion.
Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Leo Yan <leo.yan@arm.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
This patch introduces a new function cs_etm__instr_size() to calculate
the instruction size based on ISA type and instruction address.
Given the trace data can be MB and most likely that will be A64/A32 on
a lot of platforms, cs_etm__instr_addr() keeps a single ISA type check
for A64/A32 and executes an optimized calculation (addr + offset * 4).
Signed-off-by: Leo Yan <leo.yan@linaro.org>
Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Leo Yan <leo.yan@arm.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
ETE shares the same packet format as ETMv4, but exception decoding
handled ETMv4 packets only. As a result, ETE exception packets were
not classified.
Recognize the ETE magic for exception number decoding.
Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Leo Yan <leo.yan@arm.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The itrace 'c' and 'r' options request synthesized branch events for
calls and returns only. For perf script the default itrace options are
"--itrace=ce", so CS ETM should emit call branches and error events by
default.
CS ETM currently synthesizes a branch sample for every decoded taken
branch whenever branch synthesis is enabled. This produces redundant
jump and conditional branch samples.
Add a branch filter derived from the itrace calls and returns options.
When neither option is set, keep the existing behavior and synthesize all
branch samples. When calls or returns are requested, emit only branch
samples whose flags match the selected branch type, while preserving trace
begin/end markers.
Also update test_arm_coresight_disasm.sh and arm-cs-trace-disasm.py
to use the --itrace=b option for generating branch samples.
Before:
perf script -F,+flags
callchain_test 6114 [005] 331519.825214: 1 branches: tr strt jmp 0 [unknown] ([unknown]) => ffff8000803a3a68 perf_report_aux_output_id+0x50 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: call ffff8000803a3a74 perf_report_aux_output_id+0x5c ([kernel.kallsyms]) => ffff8000817f4d88 memset+0x0 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: jmp ffff8000817f4d8c memset+0x4 ([kernel.kallsyms]) => ffff8000817f4c00 __pi_memset_generic+0x0 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: jcc ffff8000817f4c1c __pi_memset_generic+0x1c ([kernel.kallsyms]) => ffff8000817f4c44 __pi_memset_generic+0x44 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: jcc ffff8000817f4c4c __pi_memset_generic+0x4c ([kernel.kallsyms]) => ffff8000817f4c5c __pi_memset_generic+0x5c ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: jcc ffff8000817f4c5c __pi_memset_generic+0x5c ([kernel.kallsyms]) => ffff8000817f4cf0 __pi_memset_generic+0xf0 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: jcc ffff8000817f4d30 __pi_memset_generic+0x130 ([kernel.kallsyms]) => ffff8000817f4d68 __pi_memset_generic+0x168 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: jcc ffff8000817f4d78 __pi_memset_generic+0x178 ([kernel.kallsyms]) => ffff8000817f4d6c __pi_memset_generic+0x16c ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: jcc ffff8000817f4d78 __pi_memset_generic+0x178 ([kernel.kallsyms]) => ffff8000817f4d6c __pi_memset_generic+0x16c ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: jcc ffff8000817f4d78 __pi_memset_generic+0x178 ([kernel.kallsyms]) => ffff8000817f4d6c __pi_memset_generic+0x16c ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: return ffff8000817f4d84 __pi_memset_generic+0x184 ([kernel.kallsyms]) => ffff8000803a3a78 perf_report_aux_output_id+0x60 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: jcc ffff8000803a3a98 perf_report_aux_output_id+0x80 ([kernel.kallsyms]) => ffff8000803a3b04 perf_report_aux_output_id+0xec ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: call ffff8000803a3b1c perf_report_aux_output_id+0x104 ([kernel.kallsyms]) => ffff8000803a38f8 __perf_event_header__init_id+0x0 ([kernel.kallsyms])
After:
callchain_test 6114 [005] 331519.825214: 1 branches: tr strt jmp 0 [unknown] ([unknown]) => ffff8000803a3a68 perf_report_aux_output_id+0x50 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: call ffff8000803a3a74 perf_report_aux_output_id+0x5c ([kernel.kallsyms]) => ffff8000817f4d88 memset+0x0 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: call ffff8000803a3b1c perf_report_aux_output_id+0x104 ([kernel.kallsyms]) => ffff8000803a38f8 __perf_event_header__init_id+0x0 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: call ffff8000803a39c0 __perf_event_header__init_id+0xc8 ([kernel.kallsyms]) => ffff800080105258 __task_pid_nr_ns+0x0 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: call ffff80008010528c __task_pid_nr_ns+0x34 ([kernel.kallsyms]) => ffff8000801d5610 __rcu_read_lock+0x0 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: call ffff8000801052b0 __task_pid_nr_ns+0x58 ([kernel.kallsyms]) => ffff800080192078 lock_acquire+0x0 ([kernel.kallsyms])
callchain_test 6114 [005] 331519.825214: 1 branches: call ffff8000801923f4 lock_acquire+0x37c ([kernel.kallsyms]) => ffff8000801d6da0 rcu_is_watching+0x0 ([kernel.kallsyms])
Fixes: b12235b113cf ("perf tools: Add mechanic to synthesise CoreSight trace packets")
Signed-off-by: Leo Yan <leo.yan@linaro.org>
Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Leo Yan <leo.yan@arm.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
cs_etm__init_traceid_queue() allocates the frontend and decode threads,
if a later allocation fails, the error path does not drop thread
reference that was already acquired.
Release both thread pointers with thread__zput() on the error path, so
does not leak thread references or leave stale pointers behind.
Fixes: 951ccccdc715 ("perf cs-etm: Only track threads instead of PID and TIDs")
Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Leo Yan <leo.yan@arm.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
It can now pass command line arguments to perf kvm record. Let's pass
'sleep 1' directly and see it doesn't fail.
$ sudo perf test -vv kvm
96: perf kvm tests:
---- start ----
test child forked, pid 3686726
Starting qemu-system-x86_64...
Testing perf kvm stat
Recording kvm events for pid 3686746 (duration 1s)...
perf kvm stat test [Success]
Testing perf kvm record/report
Recording kvm profile for pid 3686746 (duration 1s)...
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.576 MB /tmp/__perf_kvm_test.perf.data.0HgX6 (4009 samples) ]
perf kvm record/report test [Success]
Testing perf kvm buildid-list
perf kvm buildid-list test [Success]
Testing perf kvm stat live
perf kvm stat live test [Success]
Testing perf kvm record default event with command line
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.000 MB /dev/null ]
perf kvm record default event [Success]
---- end(0) ----
96: perf kvm tests : Ok
=== Test Summary ===
Passed main tests : 1
Passed subtests : 0
Skipped tests : 0
Failed tests : 0
Tested-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The first element of the kvm_usage string is allocated dynamically by
parse_options_subcommand() if it's NULL. Callers should free it after
use.
Tested-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
As it removed STRDUP_FAIL_EXIT(), it no longer calls free() for elements
in the copied argv. Thus, the filename should not be allocated as well.
In fact, it's a pointer to string literals and should be fine to pass
the pointer.
Tested-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
It's used to pass command line options to a copied argv. But there's no
reason to make the copies as it's all used in the same function. It can
simply use stack variables.
In fact, it fixes a subtle double free issue. As parse_options() can
move contents in argv[], some entries may point to the same item. So
freeing all items in the argv could trigger a double free. With stack
variables, we don't need to allocate and free them.
Tested-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
There's a subtle issue with option parsing in perf record. It calls the
function with PARSE_OPT_STOP_AT_NON_OPTION so that it can pass later
options to the external command it runs.
But perf kvm record passes the default arch events after the argv. So
if user calls it with command, then it passes the event to the external
command and fails it like below:
$ sudo perf kvm --host record sleep 1
sleep: invalid option -- 'e'
Try 'sleep --help' for more information.
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.046 MB perf.data.kvm (5 samples) ]
We can check if the default options are needed before passing the extra
command line to make sure it's passed to perf record.
Tested-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The kvm_add_default_arch_event() has a similar logic in each arch to
check if there's an existing command line option for events. Let's
check it in the generic code and remove the duplication.
Tested-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Livepatch symbols (.klp.sym.*) carry a [module] tag but resolve to core
kernel text addresses. When kcore_copy__process_kallsyms() encounters
these symbols, they are treated as module symbols, pulling the
first_module_symbol down to a kernel text address.
This corrupts the module memory range used to build the kcore PT_LOAD
segments. For example, with a kpatch module containing a
".klp.sym.vmlinux.arch_release_task_struct,0" livepatch symbol loaded:
kernel symbols
...
ffffffffb4a41120 arch_release_task_struct
... ^
... | aliased by .klp.sym
... | drags first_module_symbol here
|
(43M gap) | bloated kcore segment
|
module symbols |
ffffffffc047b000 <-- correct first_module_symbol
...
...
This causes the module PT_LOAD segment to start at the .klp.sym address
and not the real first module address, bloating the kcore copy:
Baseline (no livepatch): VirtAddr ffffffffc047b000, 8.5M
Bloated (with livepatch): VirtAddr ffffffffb4a41000, 54M
Post-fix (with livepatch): VirtAddr ffffffffc047b000, 8.9M
Filter livepatch symbols early in kcore_copy__process_kallsyms() before
they can affect module boundary tracking.
Reported-by: Sashiko <sashiko-bot@kernel.org>
Link: https://sashiko.dev/#/patchset/20260624201254.472576-1-joe.lawrence@redhat.com?part=1
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
Acked-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Livepatch modules contain special symbols (prefixed by ".klp.sym.") that
act as relocation placeholders. Once resolved, they point to the same
addresses as the original kernel symbols they reference. [1]
These special symbols confuse the 'vmlinux symtab matches kallsyms' perf
test as kallsyms may report multiple symbols sharing a single kernel
address. For example:
kallsyms (without livepatch)
----------------------------
ffffffff81a41110 T __pfx_arch_release_task_struct
> ffffffff81a41120 T arch_release_task_struct
ffffffff81a41140 T __pfx_exit_thread
ffffffff81a41150 T exit_thread
kallsyms (with livepatch loaded)
---------------------------------
ffffffff81a41110 T __pfx_arch_release_task_struct
> ffffffff81a41120 T arch_release_task_struct
ffffffff81a41140 T __pfx_exit_thread
ffffffff81a41150 T exit_thread
> ffffffff81a41120 w .klp.sym.vmlinux.arch_release_task_struct,0 [kpatch_5_14_0_570_94_1_1_3]
When perf loads kallsyms, both symbols are inserted into the symbol
table at the same address, corrupting symbol end-address calculations
and causing test failures.
Filter out symbols prefixed with ".klp.sym." when loading kallsyms, as
they alias existing kernel symbols.
Link: https://docs.kernel.org/livepatch/module-elf-format.html#livepatch-symbols [1]
Reported-and-tested-by:
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
Acked-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Build-time linker flags were not applied when linking dlfilters.
Pass LDFLAGS for consistency with the other link rules.
Signed-off-by: Trevor Allison <tallison@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The JUnit XML output correctly captures the stderr/stdout output of failed
tests inside the <failure> element. However, for skipped tests, the output
was completely discarded and the XML only received a self-closing <skipped
message="reason"/> tag.
This expands the <skipped> element to include the test's err_output when
available, which is extremely helpful for debugging why a test was skipped
(e.g. diagnosing missing prerequisites or unexpected environment states
that triggered the skip) directly from CI systems parsing the XML report.
Assisted-by: Antigravity:gemini-3.1-pro
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
Optimize the execution of the metric validation and metric listing shell
test suites:
1. `stat_metrics_values.sh`:
The Python metric validator runs the `perf bench futex hash` workload
for each validated metric relationship. Reduce the benchmark runtime
limit from `-r 2` (2 seconds) to `-r 1` (1 second). This cuts the
workload duration in half while still generating sufficient PMU events
to satisfy non-zero threshold metric validations.
2. `stat_all_metrics.sh`:
The metric checking test runs `perf stat` sequentially across all
433+ listed metrics. Change the default workload for system-wide runs
from `sleep 0.01` to `true`. This avoids the 10ms sleep delay on each
sequential metric invocation, saving over 4 seconds of total wall
time during full test suite runs.
Assisted-by: Antigravity:gemini-3.1-pro
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The lock contention analysis test suite (`lock_contention.sh`) performs a
series of 13 separate profiling checks to verify various aggregation and
filtering parameters of `perf lock contention`. Each of these checks runs
the `perf bench sched messaging` messaging benchmark as its workload.
By default, `sched messaging` runs 10 groups of 40 processes (400
processes total) generating substantial task scheduling, context
switching, and IPC message passing. When traced system-wide for lock
events, the tracing overhead (handling millions of lock acquisitions and
releases) slows execution down significantly, causing the test suite to
take over 80 seconds.
Optimize this by introducing a scaled-down messaging benchmark workload:
`perf bench sched messaging -g 1 -p`. Running 1 group (40 processes) takes
only 0.01 seconds natively (instead of 0.08 seconds), drastically reduces
the sheer volume of lock acquire/release trace events, and reduces CPU
context switching during tracing while still generating sufficient lock
events to fully exercise the BPF/record filters.
Assisted-by: Antigravity:gemini-3.1-pro
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The off-cpu profiling test suite runs multiple recording commands with a
default workload of `sleep 1` to test the off-cpu threshold configurations
(specifically, above 999ms and below 1200ms). This adds a mandatory 3.0
seconds of sleep overhead.
Optimize this by scaling down the thresholds and workload durations by a
factor of 10:
- Use `sleep 0.1` as the workload duration.
- Change the above-threshold test to use `--off-cpu-thresh 50` and `sleep
0.1`.
- Change the below-threshold test to use `--off-cpu-thresh 500` and `sleep
0.1`.
- Update the awk period check in the above-threshold test to look for a
period greater than 50,000,000 ns (50ms) instead of 999,000,000 ns
(999ms).
This reduces raw test sleep overhead from 3.0s down to 0.3s, yielding a
~2.7 second speedup for this test.
Assisted-by: Antigravity:gemini-3.1-pro
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The branch stack sampling test (test 130) runs short iteration-based
workloads to verify syscall, kernel, and trap branch stack sampling.
Specifically, `test_syscall()` and `test_kernel_branches()` run `perf
bench syscall basic` with loop counts of 8000 and 1000, and
`test_trap_eret_branches()` runs `traploop` with 1000 iterations.
Because these loop limits are extremely small, the total benchmark
runtimes last only a few milliseconds (or less). Under high load,
virtualization, or coarse sampling conditions, PMU cycle sampling fails to
capture enough samples inside the brief benchmark loops. This leads to
false negatives where the script output lacks the expected syscall,
kernel, or trap branch entries (e.g. "ERROR: Branches missing getppid[^
]*/SYSCALL/").
Fix this by increasing the workload loop counts to 100,000 across all
three test sections. Running 100,000 loops still finishes virtually
instantaneously (less than 0.1 seconds), but generates enough iterations
to guarantee robust branch stack capture.
Fixes: b55878c90ab9 ("perf test: Add test for branch stack sampling")
Assisted-by: Antigravity:gemini-3.1-pro
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The `perf stat --bpf-counters test` fails intermittently on hybrid
architectures or systems with dynamic frequency scaling (DVFS). This
happens because the test workload (`sqrtloop`) runs for a fixed 1-second
duration, and the CPU frequency can scale dynamically between idle and
maximum frequency. As the first run runs on a cold CPU and the second run
runs on a warmed-up CPU (or vice versa), the number of instructions
executed in 1 second differs by up to 2.2x, violating the comparison
tolerance.
Also, when running as root, BPF tracepoints and scheduling programs
trigger frequently. Since standard `perf stat -e instructions` measures
both user and kernel space instructions, it counts BPF helper and program
execution overheads, whereas the BPF counters themselves do not self-
measure. This introduces a large kernel-space instruction count
discrepancy between standard and BPF counters.
Fix these issues by:
1. Switching the workload to a strictly deterministic, iteration-based
workload: `awk 'BEGIN { for (i=0; i<10000000; i++) sum+=i }'`. We pin
the
workload to a single random allowed CPU using `taskset -c $CPU` via a
bash array.
2. Restricting the counted event to user-space only (`instructions:u` or
`/u`).
3. Tightening the comparison tolerance from 20% to 15%.
These modifications isolate the measurements to user-space instructions of
the deterministic loop, which executes a virtually identical number of
instructions on both runs (with less than 0.001% variation), eliminating
Dynamic Frequency Scaling (DVFS), kernel scheduling noise, and BPF helper
self-measurement overheads.
Fixes: 2c0cb9f56020 ("perf test: Add a shell test for 'perf stat --bpf-counters' new option")
Assisted-by: Antigravity:gemini-3.1-pro
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
|
The `perf trace record and replay` test fails intermittently on slow or
virtualized hosts because the default recording workload (`sleep 1`)
occasionally completes without scheduling the target `nanosleep` or
`clock_nanosleep` system calls inside the recorded sample window,
resulting in the error: `Failed: cannot find *nanosleep syscall`.
Generalize the `perf_record_with_retry` helper in
`tests/shell/lib/perf_record.sh` to support a custom record command prefix
via the `PERF_RECORD_CMD` environment variable (defaulting to "perf
record").
Update `trace_record_replay.sh` to use this robust retry loop running with
`PERF_RECORD_CMD="perf trace record"` and a base workload of `sleep`. The
test will automatically retry with scaled sleep durations (from 0.01s up
to 2.0s) until the required `nanosleep` event is successfully captured.
Fixes: 15bcfb96d0dd ("perf test: Add trace record and replay test")
Assisted-by: Antigravity:gemini-3.1-pro
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
|
