Try to remove some, but not all duplicate toolchain definitions. In
these instances, their makefiles already include
tools/scripts/Makefile.include which defines these in a consistent way.
STRIP is the only one that was set with an '=', but I don't think it
was significant so that difference can be dropped.
Signed-off-by: James Clark <james.clark@linaro.org>
Reviewed-by: Ian Rogers <irogers@google.com>
Acked-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
If the realloc for fda->entries succeeds but the realloc for fda->priv
fails, the error path frees the newly allocated entries.
However, fda->entries is neither updated to point to the new entries block
nor cleared to NULL. If realloc moved the allocation to a new block, the
old fda->entries pointer is now freed memory.
When fdarray__exit() is later called to clean up, it executes
free(fda->entries), which would trigger a double-free on that old pointer.
Reported-by: Sashiko Review <sashiko-bot@kernel.org>
Closes: https://lore.kernel.org/linux-perf-users/20260710200150.11FE71F00A3A@smtp.kernel.org
Reviewed-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: f94563fac2 ("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>
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 59b4412f27 ("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: 59b4412f27 ("libperf: Avoid internal moving of fdarray fds")
Fixes: 657ee55319 ("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>
mount_overload() builds an environment variable name like
"PERF_SYSFS_ENVIRONMENT" from fs->name. Two bugs:
1) snprintf() uses name_len as the buffer size instead of sizeof(upper_name).
For fs->name = "sysfs" (len=5), the output is truncated to "PERF" (4
chars + null), so getenv() never finds the intended variable.
2) mem_toupper() only uppercases name_len bytes, converting just the "PERF"
prefix rather than the full string including the filesystem name portion.
Fix by using sizeof(upper_name) for snprintf and strlen(upper_name) for
mem_toupper, so the full "PERF_SYSFS_ENVIRONMENT" string is correctly
formatted and uppercased.
Reported-by: sashiko-bot <sashiko-bot@kernel.org>
Fixes: 73ca85ad36 ("tools lib api fs: Add FSTYPE__mount() method")
Cc: Jiri Olsa <jolsa@kernel.org>
Assisted-by: Claude:claude-opus-4.6
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
filename__write_int() formats an integer into a 64-byte buffer with
sprintf() then passes sizeof(buf) (64) as the write length. This
writes all 64 bytes including uninitialized stack data past the
formatted string. Most sysfs files reject the oversized write,
making the function always return -1.
Fix by capturing the sprintf() return value and using it as the
write length.
Reported-by: sashiko-bot <sashiko-bot@kernel.org>
Fixes: 3b00ea9386 ("tools lib api fs: Add sysfs__write_int function")
Cc: Kan Liang <kan.liang@intel.com>
Assisted-by: Claude:claude-opus-4.6
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
filename__read_int() passes a stack buffer to read() using the full
sizeof(line) and then hands it to atoi() without null-terminating.
If a sysfs file fills the 64-byte buffer exactly, atoi() reads past
the array into uninitialized stack memory.
filename__read_ull_base() has the same issue with strtoull().
Fix both by reading sizeof(line) - 1 bytes and explicitly
null-terminating after a successful read.
Fixes: 3a351127cb ("tools lib fs: Adopt filename__read_int from tools/perf/")
Reported-by: sashiko-bot <sashiko-bot@kernel.org>
Assisted-by: Claude:claude-opus-4.6
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
glibc's opendir allocates a minimum of 32kb, when called recursively
for a directory tree the memory consumption can add up - nearly 300kb
during perf start-up when processing modules. Add a stack allocated
variant of readdir sized a little more than 1kb.
As getdents64 may be missing from libc, add support using syscall. As
the system call number maybe missing, add #defines for those.
Note, an earlier version of this patch had a feature test for
getdents64 but there were problems on certains distros where
getdents64 would be #define renamed to getdents breaking the code. The
syscall use was made uncondtional to work around this. There is
context in:
https://lore.kernel.org/lkml/20231207050433.1426834-1-irogers@google.com/
Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250222061015.303622-2-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
When executing the command "perf list", I met "Error: failed to open
tracing events directory" twice, the first reason is that there is no
"/sys/kernel/tracing/events" directory due to it does not enable the
kernel tracing infrastructure with CONFIG_FTRACE, the second reason
is that there is no root privileges.
Add the error string to tell the users what happened and what should
to do, and also call put_tracing_file() to free events_path a little
later to avoid messy code in the error message.
At the same time, just remove the redundant "/" of the file path in
the function get_tracing_file(), otherwise it shows something like
"/sys/kernel/tracing//events".
Before:
$ ./perf list
Error: failed to open tracing events directory
After:
(1) Without CONFIG_FTRACE
$ ./perf list
Error: failed to open tracing events directory
/sys/kernel/tracing/events: No such file or directory
(2) With CONFIG_FTRACE but no root privileges
$ ./perf list
Error: failed to open tracing events directory
/sys/kernel/tracing/events: Permission denied
Committer testing:
Redirect stdout to null to quickly test the patch:
Before:
$ perf list > /dev/null
Error: failed to open tracing events directory
$
After:
$ perf list > /dev/null
Error: failed to open tracing events directory
/sys/kernel/tracing/events: Permission denied
$
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/lkml/20240730062301.23244-3-yangtiezhu@loongson.cn
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
In general a read fills 4kb so filling the buffer is a 1 in 4096
operation, move it out of the io__get_char function to avoid some
checking overhead and to better hint the function is good to inline.
For perf's IO intensive internal (non-rigorous) benchmarks there's a
small improvement to kallsyms-parsing with a default build.
Before:
```
$ perf bench internals all
Computing performance of single threaded perf event synthesis by
synthesizing events on the perf process itself:
Average synthesis took: 146.322 usec (+- 0.305 usec)
Average num. events: 61.000 (+- 0.000)
Average time per event 2.399 usec
Average data synthesis took: 145.056 usec (+- 0.155 usec)
Average num. events: 329.000 (+- 0.000)
Average time per event 0.441 usec
Average kallsyms__parse took: 162.313 ms (+- 0.599 ms)
...
Computing performance of sysfs PMU event scan for 100 times
Average core PMU scanning took: 53.720 usec (+- 7.823 usec)
Average PMU scanning took: 375.145 usec (+- 23.974 usec)
```
After:
```
$ perf bench internals all
Computing performance of single threaded perf event synthesis by
synthesizing events on the perf process itself:
Average synthesis took: 127.829 usec (+- 0.079 usec)
Average num. events: 61.000 (+- 0.000)
Average time per event 2.096 usec
Average data synthesis took: 133.652 usec (+- 0.101 usec)
Average num. events: 327.000 (+- 0.000)
Average time per event 0.409 usec
Average kallsyms__parse took: 150.415 ms (+- 0.313 ms)
...
Computing performance of sysfs PMU event scan for 100 times
Average core PMU scanning took: 47.790 usec (+- 1.178 usec)
Average PMU scanning took: 376.945 usec (+- 23.683 usec)
```
Signed-off-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20240519181716.4088459-1-irogers@google.com
Multiple threads, such as with "perf top", may race to initialize a
file system path like hugetlbfs. The racy initialization of the path
leads to at least memory leaks. To avoid this initialize each fs for
reading the mount point path with pthread_once.
Mounting the file system may also be racy, so introduce a mutex over
the function. This does mean that the path is being accessed with and
without a mutex, which is inherently racy but hopefully benign,
especially as there are fewer callers to fs__mount.
Remove the fs__entries by directly using global variables, this was
done as no argument like the index can be passed to the init once
routine.
Issue found and tested with "perf top" and address sanitizer.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: bpf@vger.kernel.org
Link: https://lore.kernel.org/r/20230609224004.180988-1-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The canonical location for the tracefs filesystem is at /sys/kernel/tracing.
But, from Documentation/trace/ftrace.rst:
Before 4.1, all ftrace tracing control files were within the debugfs
file system, which is typically located at /sys/kernel/debug/tracing.
For backward compatibility, when mounting the debugfs file system,
the tracefs file system will be automatically mounted at:
/sys/kernel/debug/tracing
Many comments and Kconfig help messages in the tracing code still refer
to this older debugfs path, so let's update them to avoid confusion.
Link: https://lore.kernel.org/linux-trace-kernel/20230215223350.2658616-2-zwisler@google.com
Acked-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Reviewed-by: Mukesh Ojha <quic_mojha@quicinc.com>
Signed-off-by: Ross Zwisler <zwisler@google.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
