linux-toradex.git/tools/perf/util/Build, branch master

perf kvm stat: Remove use of the arch directory

2026-02-03T21:01:27+00:00

`perf kvm stat` supports record and report options.

By using the arch directory a report for a different machine type cannot
be supported.

Move the kvm-stat code out of the arch directory and into
util/kvm-stat-arch following the pattern of perf-regs and dwarf-regs.

Avoid duplicate symbols by renaming functions to have the architecture
name within them.

For global variables, wrap them in an architecture specific function.
Selecting the architecture to use with `perf kvm stat` is selected by
EM_HOST, ie no different than before the change.

Later the ELF machine can be determined from the session or a header
feature (ie EM_HOST at the time of the record).

The build and #define HAVE_KVM_STAT_SUPPORT is now redundant so remove
across Makefiles and in the build.

Opportunistically constify architectural structs and arrays.

Signed-off-by: Ian Rogers 
Cc: Aditya Bodkhe 
Cc: Adrian Hunter 
Cc: Albert Ou 
Cc: Alexander Shishkin 
Cc: Alexandre Ghiti 
Cc: Andi Kleen 
Cc: Andrew Jones 
Cc: Anubhav Shelat 
Cc: Anup Patel 
Cc: Athira Rajeev 
Cc: Blake Jones 
Cc: Chun-Tse Shao 
Cc: Dapeng Mi 
Cc: Dmitriy Vyukov 
Cc: Howard Chu 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Leo Yan 
Cc: Namhyung Kim 
Cc: Palmer Dabbelt 
Cc: Paul Walmsley 
Cc: Peter Zijlstra 
Cc: Quan Zhou 
Cc: Shimin Guo 
Cc: Swapnil Sapkal 
Cc: Thomas Falcon 
Cc: Will Deacon 
Cc: Yunseong Kim 
Signed-off-by: Arnaldo Carvalho de Melo

perf capstone: Support for dlopen-ing libcapstone.so

2026-02-03T14:34:02+00:00

If perf is built with LIBCAPSTONE_DLOPEN=1, support dlopen-ing
libcapstone.so and then calling the necessary functions by looking them
up using dlsym.

The types come from capstone.h which means the libcapstone feature check
needs to pass, and NO_CAPSTONE=1 hasn't been defined. This will cause
the definition of HAVE_LIBCAPSTONE_SUPPORT.

Earlier versions of this code tried to declare the necessary
capstone.h constants and structs, but they weren't stable and caused
breakages across libcapstone releases.

Signed-off-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Bill Wendling 
Cc: Charlie Jenkins 
Cc: Collin Funk 
Cc: Dmitriy Vyukov 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: Justin Stitt 
Cc: Masami Hiramatsu 
Cc: Namhyung Kim 
Cc: Nathan Chancellor 
Cc: Nick Desaulniers 
Cc: Peter Zijlstra 
Signed-off-by: Arnaldo Carvalho de Melo

perf disasm: Don't include C files from the arch directory

2026-01-23T19:58:38+00:00

Move the arch instructions.c files into appropriately named files in
annotate-arch in the util directory.

Don't #include to compile the code, switch to building the files and fix
up the #includes accordingly.

Move powerpc specific disasm code out of disasm.c and into
annotate-powerpc.c.

Declarations and static removed as appropriate for the code to compile
as separate compilation units.

The e_machine and e_flags set up is moved to the disasm.c architectures
array so that later patches can sort by them.

Reviewed-by: James Clark 
Signed-off-by: Ian Rogers 
Cc: Aditya Bodkhe 
Cc: Adrian Hunter 
Cc: Albert Ou 
Cc: Alexander Shishkin 
Cc: Alexandre Ghiti 
Cc: Athira Rajeev 
Cc: Bill Wendling 
Cc: Dr. David Alan Gilbert 
Cc: Guo Ren 
Cc: Howard Chu 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Julia Lawall 
Cc: Justin Stitt 
Cc: Krzysztof Łopatowski 
Cc: Leo Yan 
Cc: Namhyung Kim 
Cc: Nathan Chancellor 
Cc: Nick Desaulniers 
Cc: Palmer Dabbelt 
Cc: Paul Walmsley 
Cc: Peter Zijlstra 
Cc: Sergei Trofimovich 
Cc: Shimin Guo 
Cc: Suchit Karunakaran 
Cc: Thomas Falcon 
Cc: Tianyou Li 
Cc: Will Deacon 
Cc: Zecheng Li 
Signed-off-by: Arnaldo Carvalho de Melo

perf build: Remove NO_LIBDW_DWARF_UNWIND option

2026-01-20T18:43:56+00:00

Libdw unwinding support is present for every architecture that has a
perf_regs.h - perf registers are needed for the initial frame to
unwind.

Elfutils also supports SPARC, ARC and m68k but there is no support in
the Linux kernel for perf registers on these architectures.

As the perf supported DWARF unwinding architectures are a subset of the
elfutils ones, remove NO_LIBDW_DWARF_UNWIND as there isn't a case of
elfutils lacking the support need for perf.

Signed-off-by: Ian Rogers 
Cc: Aditya Bodkhe 
Cc: Adrian Hunter 
Cc: Albert Ou 
Cc: Alexandre Ghiti 
Cc: Andi Kleen 
Cc: Athira Rajeev 
Cc: Chun-Tse Shao 
Cc: Dmitriy Vyukov 
Cc: Dr. David Alan Gilbert 
Cc: Guo Ren 
Cc: Haibo Xu 
Cc: Howard Chu 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Krzysztof Łopatowski 
Cc: Leo Yan 
Cc: Mark Wielaard 
Cc: Namhyung Kim 
Cc: Palmer Dabbelt 
Cc: Paul Walmsley 
Cc: Peter Zijlstra 
Cc: Sergei Trofimovich 
Cc: Shimin Guo 
Cc: Stephen Brennan 
Cc: Thomas Falcon 
Cc: Will Deacon 
Signed-off-by: Arnaldo Carvalho de Melo

perf dwarf-regs: Add util/dwarf-regs-arch for consistency with perf-regs

2026-01-20T18:43:56+00:00

perf_regs.h has cross architecture functions for operating with the
differing perf register constants. dwarf-regs.h is similar but for
cross architecture dwarf notions of registers.

For consistency move the arch parts of dwarf-regs out of util and into
its own directory.

Signed-off-by: Ian Rogers 
Cc: Aditya Bodkhe 
Cc: Adrian Hunter 
Cc: Albert Ou 
Cc: Alexandre Ghiti 
Cc: Andi Kleen 
Cc: Athira Rajeev 
Cc: Chun-Tse Shao 
Cc: Dmitriy Vyukov 
Cc: Dr. David Alan Gilbert 
Cc: Guo Ren 
Cc: Haibo Xu 
Cc: Howard Chu 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Krzysztof Łopatowski 
Cc: Leo Yan 
Cc: Mark Wielaard 
Cc: Namhyung Kim 
Cc: Palmer Dabbelt 
Cc: Paul Walmsley 
Cc: Peter Zijlstra 
Cc: Sergei Trofimovich 
Cc: Shimin Guo 
Cc: Stephen Brennan 
Cc: Thomas Falcon 
Cc: Will Deacon 
Signed-off-by: Arnaldo Carvalho de Melo

perf unwind-libdw: Fix a cross-arch unwinding bug

2026-01-20T16:00:39+00:00

The set_initial_registers field of Dwfl_Thread_Callbacks needs to be set
according to the arch of the stack samples being analyzed, not the arch
that perf itself is built for.

Currently perf fails to unwind stack samples collected from archs
different from that of the host perf is running on.

This patch moves the arch-specific implementations of set_initial_registers
from tools/perf/arch to tools/perf/utli/unwind-libdw-arch, similar to the
way the perf-regs-arch folder contains arch-specific functions related to
registers, and chooses the implementation based on the arch of the data
being processed.

Reviewed-by: Ian Rogers 
Signed-off-by: Shimin Guo 
Acked-by: Namhyung Kim 
Cc: Aditya Bodkhe 
Cc: Adrian Hunter 
Cc: Albert Ou 
Cc: Alexandre Ghiti 
Cc: Andi Kleen 
Cc: Athira Rajeev 
Cc: Chun-Tse Shao 
Cc: Dmitriy Vyukov 
Cc: Dr. David Alan Gilbert 
Cc: Guo Ren 
Cc: Haibo Xu 
Cc: Howard Chu 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Krzysztof Łopatowski 
Cc: Leo Yan 
Cc: Mark Wielaard 
Cc: Palmer Dabbelt 
Cc: Paul Walmsley 
Cc: Peter Zijlstra 
Cc: Sergei Trofimovich 
Cc: Stephen Brennan 
Cc: Thomas Falcon 
Cc: Will Deacon 
Signed-off-by: Ian Rogers 
Signed-off-by: Arnaldo Carvalho de Melo

perf help: Move common_cmds into builtin-help

2026-01-14T20:22:50+00:00

There's a lot of infrastructure for generating a relatively simple
array used by one function.

Move the array into the function and remove the supporting build logic.

At the same time opportunistically const-ify the array.

Signed-off-by: Ian Rogers 
Acked-by: Namhyung Kim 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Charlie Jenkins 
Cc: Howard Chu 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: Peter Zijlstra 
Signed-off-by: Arnaldo Carvalho de Melo

perf util: Remove SHA-1 code

2026-01-13T20:03:32+00:00

Now that the SHA-1 code is no longer used, remove it.

Signed-off-by: Eric Biggers 
Tested-by: Ian Rogers 
Acked-by: Namhyung Kim 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Fangrui Song 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jason A. Donenfeld 
Cc: Jiri Olsa 
Cc: Mark Rutland 
Cc: Pablo Galindo 
Cc: Peter Zijlstra 
Signed-off-by: Arnaldo Carvalho de Melo

perf util: Add BLAKE2s support

2026-01-13T20:03:25+00:00

Add BLAKE2s support to the perf utility library.  The code is borrowed
from the kernel.  This will replace the use of SHA-1 in genelf.c.

Signed-off-by: Eric Biggers 
Tested-by: Ian Rogers 
Acked-by: Namhyung Kim 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Fangrui Song 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jason A. Donenfeld 
Cc: Jiri Olsa 
Cc: Mark Rutland 
Cc: Pablo Galindo 
Cc: Peter Zijlstra 
Signed-off-by: Arnaldo Carvalho de Melo

perf addr2line: Add a libdw implementation

2026-01-12T19:23:06+00:00

Add an implementation of addr2line that uses libdw.

Other addr2line implementations are slow, particularly in the case of
forking addr2line.

Add an implementation that caches the libdw information in the dso and
uses it to find the file and line number information.

Inline information is supported but because cu_walk_functions_at visits
the leaf function last add a inline_list__append_tail to reverse the
lists order.

Committer testing:

  # perf probe -x ~/bin/perf libdw__addr2line
  Added new event:
    probe_perf:libdw_addr2line (on libdw__addr2line in /home/acme/bin/perf)

  You can now use it in all perf tools, such as:

  	perf record -e probe_perf:libdw_addr2line -aR sleep 1

  #
  # perf stat -e probe_perf:libdw_addr2line perf report -f --dso perf --stdio -s srcfile,srcline
  # To display the perf.data header info, please use --header/--header-only options.
  #
  #
  # Total Lost Samples: 0
  #
  # Samples: 4K of event 'cpu/cycles/Pu'
  # Event count (approx.): 5535180842
  #
  # Overhead  Source File   Source:Line
  # ........  ............  ...............
  #
      99.04%  inlineloop.c  inlineloop.c:21
       0.46%  inlineloop.c  inlineloop.c:20

  #
  # (Tip: For tracepoint events, try: perf report -s trace_fields)
  #

   Performance counter stats for 'perf report -f --dso perf --stdio -s srcfile,srcline':

                  44      probe_perf:libdw_addr2line

         0.037260744 seconds time elapsed

         0.025299000 seconds user
         0.011918000 seconds sys
  #

Adding probes to the other addr2line implementations (llvm__addr2line,
libbfd__addr2line and cmd__addr2line) I noticed some fallbacks to the
llvm one:

 Performance counter stats for 'perf report -f --dso perf --stdio -s srcfile,srcline':

                44      probe_perf:libdw_addr2line
                23      probe_perf:llvm_addr2line
                 0	probe_perf:libbfd_addr2line
                 0      probe_perf:cmd_addr2line

Something to investigate further, but at least we don't fallback to the
cmd based one :-)

Reviewed-by: James Clark 
Signed-off-by: Ian Rogers 
Tested-by: Arnaldo Carvalho de Melo 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Howard Chu 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Stephen Brennan 
Cc: Tony Jones 
Signed-off-by: Arnaldo Carvalho de Melo