linux-stable.git/tools/perf, branch v6.18

perf libbfd: Ensure libbfd is initialized prior to use

2025-11-13T20:55:19+00:00

Multiple threads may be creating and destroying BFD objects in
situations like `perf top`.

Without appropriate initialization crashes may occur during libbfd's
cache management.

BFD's locks require recursive mutexes, add support for these.

Committer testing:

This happens only when building with 'make BUILD_NONDISTRO=1' and having
the binutils-devel package (or equivalent) installed, i.e. linking with
binutils devel files, an opt-in perf build.

Before:

  root@x1:~# perf top
  perf: Segmentation fault
  -------- backtrace --------
  
  root@x1:~#

After this patch it works as before.

Closes: https://lore.kernel.org/lkml/aQt66zhfxSA80xwt@gentoo.org/
Fixes: 95931d9a594dd0b5 ("perf libbfd: Move libbfd functionality to its own file")
Reported-by: Guilherme Amadio 
Signed-off-by: Ian Rogers 
Tested-by: Arnaldo Carvalho de Melo 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Signed-off-by: Arnaldo Carvalho de Melo

perf test: Fix lock contention test

2025-11-13T20:29:00+00:00

Couple of independent fixes:

1. Wire in SIGSEGV handler that terminates the test with a failure code.

2. Use "--lock-cgroup" instead of "-g"; "-g" was proposed but never
   merged. See commit 4d1792d0a2564caf ("perf lock contention: Add
   --lock-cgroup option")

3. Call cleanup() on every normal exit so trap_cleanup() doesn't mistake
   it for an unexpected signal and emit a false-negative "Unexpected
   signal in main" message.

Before patch:

  # ./perf test -vv "lock contention"
   85: kernel lock contention analysis test:
  --- start ---
  test child forked, pid 610711
  Testing perf lock record and perf lock contention
  Testing perf lock contention --use-bpf
  Testing perf lock record and perf lock contention at the same time
  Testing perf lock contention --threads
  Testing perf lock contention --lock-addr
  Testing perf lock contention --lock-cgroup
  Unexpected signal in test_aggr_cgroup
  ---- end(0) ----
   85: kernel lock contention analysis test                            : Ok

After patch:

  # ./perf test -vv "lock contention"
   85: kernel lock contention analysis test:
  --- start ---
  test child forked, pid 602637
  Testing perf lock record and perf lock contention
  Testing perf lock contention --use-bpf
  Testing perf lock record and perf lock contention at the same time
  Testing perf lock contention --threads
  Testing perf lock contention --lock-addr
  Testing perf lock contention --lock-cgroup
  Testing perf lock contention --type-filter (w/ spinlock)
  Testing perf lock contention --lock-filter (w/ tasklist_lock)
  Testing perf lock contention --callstack-filter (w/ unix_stream)
  [Skip] Could not find 'unix_stream'
  Testing perf lock contention --callstack-filter with task aggregation
  [Skip] Could not find 'unix_stream'
  Testing perf lock contention --cgroup-filter
  Testing perf lock contention CSV output
  ---- end(0) ----
   85: kernel lock contention analysis test                            : Ok

Reviewed-by: Ian Rogers 
Signed-off-by: Ravi Bangoria 
Tested-by: Arnaldo Carvalho de Melo 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Ananth Narayan 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Sandipan Das 
Cc: Santosh Shukla 
Cc: Tycho Andersen 
Signed-off-by: Arnaldo Carvalho de Melo

perf lock: Fix segfault due to missing kernel map

2025-11-13T20:17:41+00:00

Kernel maps are encoded in PERF_RECORD_MMAP2 samples but "perf lock
report" and "perf lock contention" do not process MMAP2 samples.

Because of that, machine->vmlinux_map stays NULL and any later access
triggers a segmentation fault.

Fix it by adding ->mmap2() callbacks.

Fixes: 53b00ff358dc75b1 ("perf record: Make --buildid-mmap the default")
Reported-by: Tycho Andersen (AMD) 
Reviewed-by: Ian Rogers 
Signed-off-by: Ravi Bangoria 
Tested-by: Tycho Andersen (AMD) 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Ananth Narayan 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Sandipan Das 
Cc: Santosh Shukla 
Signed-off-by: Arnaldo Carvalho de Melo

perf build: Don't fail fast path feature detection when binutils-devel is not available

2025-11-13T20:16:34+00:00

This is one more remnant of the BUILD_NONDISTRO series to make building
with binutils-devel opt-in due to license incompatibility.

In this case just the references at link time were still in place, which
make building the test-all.bin file fail, which wasn't detected before
probably because the last test was done with binutils-devel available,
doh.

Now:

  $ rpm -q binutils-devel
  package binutils-devel is not installed
  $ file /tmp/build/perf-tools/feature/test-all.bin
  /tmp/build/perf-tools/feature/test-all.bin: ELF 64-bit LSB executable, x86-64, version 1 (SYSV),
  dynamically linked, interpreter /lib64/ld-linux-x86-64.so.2,
  BuildID[sha1]=4b5388a346b51f1b993f0b0dbd49f4570769b03c, for GNU/Linux 3.2.0, not stripped
  $

Fixes: 970ae86307718c34 ("perf build: The bfd features are opt-in, stop testing for them by default")
Reviewed-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Signed-off-by: Arnaldo Carvalho de Melo

perf header: Write bpf_prog (infos|btfs)_cnt to data file

2025-11-13T20:16:33+00:00

With commit f0d0f978f3f5830a ("perf header: Don't write empty BPF/BTF
info"), the write_bpf_( prog_info() | btf() ) functions exit without
writing anything if env->bpf_prog.(infos| btfs)_cnt is zero.

process_bpf_( prog_info() | btf() ), however, still expect a "count"
value to exist in the data file. If btf information is empty, for
example, process_bpf_btf will read garbage or some other data as the
number of btf nodes in the data file. As a result, the data file will
not be processed correctly.

Instead, write the count to the data file and exit if it is zero.

Fixes: f0d0f978f3f5830a ("perf header: Don't write empty BPF/BTF info")
Reviewed-by: Ian Rogers 
Signed-off-by: Thomas Falcon 
Acked-by: Namhyung Kim 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Mark Rutland 
Cc: Peter Zijlstra 
Signed-off-by: Arnaldo Carvalho de Melo

perf symbols: Handle '1' symbols in /proc/kallsyms

2025-11-03T17:54:31+00:00

I started seeing this in recent Fedora 42 kernels:

  root@x1:~# uname -a
  Linux x1 6.17.4-200.fc42.x86_64 #1 SMP PREEMPT_DYNAMIC Sun Oct 19 18:47:49 UTC 2025 x86_64 GNU/Linux
  root@x1:~#

  root@x1:~# perf test 1
    1: vmlinux symtab matches kallsyms     : FAILED!
  root@x1:~#

Related to:

  root@x1:~# grep ' 1 ' /proc/kallsyms
  ffffffffb098bc00 1 __pfx__RNCINvNtNtNtCsfwaGRd4cjqE_4core4iter8adapters3map12map_try_foldjNtCskFudTml27HW_12drm_panic_qr7VersionuINtNtNtBa_3ops12control_flow11ControlFlowB10_ENcB10_0NCINvNvNtNtNtB8_6traits8iterator8Iterator4find5checkB10_NCNvMB12_B10_13from_segments0E0E0B12_
  ffffffffb098bc10 1 _RNCINvNtNtNtCsfwaGRd4cjqE_4core4iter8adapters3map12map_try_foldjNtCskFudTml27HW_12drm_panic_qr7VersionuINtNtNtBa_3ops12control_flow11ControlFlowB10_ENcB10_0NCINvNvNtNtNtB8_6traits8iterator8Iterator4find5checkB10_NCNvMB12_B10_13from_segments0E0E0B12_
  root@x1:~#

That is found in:

  root@x1:~# pahole --running_kernel_vmlinux
  /usr/lib/debug/lib/modules/6.17.4-200.fc42.x86_64/vmlinux
  root@x1:~#

  root@x1:~# readelf -sW /usr/lib/debug/lib/modules/6.17.4-200.fc42.x86_64/vmlinux | grep __pfx__RNCINvNtNtNtCsfwaGRd4cjqE_4core4iter8adapters3map12map_try_foldjNtCskFudTml27HW_12drm_panic_qr7VersionuINtNtNtBa_3ops12control_flow11ControlFlowB10_ENcB10_0NCINvNvNtNtNtB8_6traits8iterator8Iterator4find5checkB10_NCNvMB12_B10_13from_segments0E0E0B12_
  150649: ffffffff81f8bc00    16 FUNC    LOCAL  DEFAULT    1 __pfx__RNCINvNtNtNtCsfwaGRd4cjqE_4core4iter8adapters3map12map_try_foldjNtCskFudTml27HW_12drm_panic_qr7VersionuINtNtNtBa_3ops12control_flow11ControlFlowB10_ENcB10_0NCINvNvNtNtNtB8_6traits8iterator8Iterator4find5checkB10_NCNvMB12_B10_13from_segments0E0E0B12_
  root@x1:~#

But was being filtered out when reading /proc/kallsyms, as the '1'
symbol type was not being handled, do it, there are just two of them at
this point.

Cc: Alex Gaynor 
Cc: Alice Ryhl 
Cc: Andreas Hindborg 
Cc: Benno Lossin 
Cc: Björn Roy Baron 
Cc: Boqun Feng 
Cc: Danilo Krummrich 
Cc: Gary Guo 
Cc: Miguel Ojeda 
Cc: Trevor Gross 
Signed-off-by: Arnaldo Carvalho de Melo

tools headers x86: Sync table due to introducion of uprobe syscall

2025-11-01T16:04:20+00:00

To pick the changes in this cset:

  56101b69c9190667 ("uprobes/x86: Add uprobe syscall to speed up uprobe")

That add support for this new 'uprobe' syscall in tools such as 'perf trace'.

Now it is possible to do a system wide 'perf trace' to look if this new
syscall is being used:

  root@number:~# perf trace -v -e uprobe
  
  event qualifier tracepoint filter: (common_pid != 33989) && (id == 336)
  ^C
  root@number#

  $ grep -w uprobe tools/perf/arch/x86/entry/syscalls/syscall_64.tbl
  336	common	uprobe			sys_uprobe
  $

This addresses these perf build warnings:

  Warning: Kernel ABI header differences:
    diff -u tools/perf/arch/x86/entry/syscalls/syscall_64.tbl arch/x86/entry/syscalls/syscall_64.tbl

Please see tools/include/uapi/README for further details.

Cc: Jiri Olsa 
Cc: Peter Zijlstra 
Signed-off-by: Arnaldo Carvalho de Melo

tools headers: Sync uapi/linux/fcntl.h with the kernel sources

2025-11-01T16:01:13+00:00

To pick up the changes in this cset:

  e83f0b5d10dcf628 ("nsfs: support exhaustive file handles")

That doesn't introduce anything of interest for tools/, just addresses
these perf build warnings:

Warning: Kernel ABI header differences:
  diff -u tools/perf/trace/beauty/include/uapi/linux/fcntl.h include/uapi/linux/fcntl.h

Please see tools/include/uapi/README for further details.

Cc: Christian Brauner 
Signed-off-by: Arnaldo Carvalho de Melo

tools headers: Sync uapi/linux/prctl.h with the kernel source

2025-11-01T15:53:56+00:00

To pick up the changes in these csets:

  8cdc4d27019356b0 ("mm/huge_memory: respect MADV_COLLAPSE with PR_THP_DISABLE_EXCEPT_ADVISED")
  9dc21bbd62edeae6 ("prctl: extend PR_SET_THP_DISABLE to optionally exclude VM_HUGEPAGE")

That don't introduce anything of interest for the tools/, just
addressing these perf build warnings:

  Warning: Kernel ABI header differences:
    diff -u tools/perf/trace/beauty/include/uapi/linux/prctl.h include/uapi/linux/prctl.h

Please see tools/include/uapi/README for further details.

Cc: Andrew Morton 
Cc: David Hildenbrand 
Signed-off-by: Arnaldo Carvalho de Melo

tools headers uapi: Update fs.h with the kernel sources

2025-11-01T15:24:27+00:00

To pick up changes from:

  db2ab24a341ce893 ("Add RWF_NOSIGNAL flag for pwritev2")

These are used to beautify fs syscall arguments, albeit the changes in
this update are not affecting those beautifiers.

This addresses these tools/ build warnings:

  Warning: Kernel ABI header differences:
  diff -u tools/perf/trace/beauty/include/uapi/linux/fs.h include/uapi/linux/fs.h

Please see tools/include/uapi/README for details (it's in the first patch
of this series).

Cc: Adrian Hunter 
Cc: Christian Brauner 
Cc: Ian Rogers 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: Kan Liang 
Cc: Lauri Vasama 
Cc: Namhyung Kim 
Signed-off-by: Arnaldo Carvalho de Melo