linux-toradex.git/tools/perf/arch/powerpc, branch master

perf dwarf: Avoid redefinition warnings for REG_DWARFNUM_NAME

2026-06-10T19:53:35+00:00

dwarf-regs.c includes an arch-specific dwarf-regs-table.h for several
architectures. This pulls in different definitions of REG_DWARFNUM_NAME
and causes compiler warnings for W=1 builds.

In file included from util/dwarf-regs.c:23:

  .../dwarf-regs-table.h:5: error: "REG_DWARFNUM_NAME" redefined [-Werror]
   #define REG_DWARFNUM_NAME(reg, idx) [idx] = reg

Undefine REG_DWARFNUM_NAME before each new definition.

Suggested-by: Arnaldo Carvalho de Melo 
Signed-off-by: Martin Kaiser 
Cc: Ian Rogers 
Cc: Namhyung Kim 
Signed-off-by: Arnaldo Carvalho de Melo

powerpc tools perf: Initialize error code in auxtrace_record_init function

2026-05-27T11:08:53+00:00

perf trace record fails some cases in powerpc

 # perf test "perf trace record and replay"
 128: perf trace record and replay                                    : FAILED!

  # perf trace record sleep 1
  # echo $?
    32

This is happening because of non-zero err value from
auxtrace_record__init() function.

 static int record__auxtrace_init(struct record *rec)
 {
        int err;

        if ((rec->opts.auxtrace_snapshot_opts || rec->opts.auxtrace_sample_opts)
            && record__threads_enabled(rec)) {
                pr_err("AUX area tracing options are not available in parallel streaming mode.\n");
                return -EINVAL;
        }

        if (!rec->itr) {
                rec->itr = auxtrace_record__init(rec->evlist, &err);
                if (err)
                        return err;
        }

Here "int err" is not initialised. The code expects "err" to be set from
auxtrace_record__init() function.

Update auxtrace_record__init() in arch/powerpc/util/auxtrace.c to clear
err value in the beginning.

- Clear err value in beginning of function. Any fail later will
set appropriate return code to err.

- Even if we haven't found any event for auxtrace, perf record
should continue for other events. NULL return
will indicate that there is no auxtrace record initialized.

- Not having "err" set here will affect monitoring of other events
also because perf record will fail seeing random value in err.

Set err to -EINVAL before invoking auxtrace_record__init() in
builtin-record.c

With the fix,

  # perf trace record sleep 1
  [ perf record: Woken up 2 times to write data ]
  [ perf record: Captured and wrote 0.033 MB perf.data (228 samples) ]

Fixes: 1dbfaf94cf66ec4b ("perf powerpc: Add basic CONFIG_AUXTRACE support for VPA pmu on powerpc")
Reviewed-by: Adrian Hunter 
Signed-off-by: Athira Rajeev 
Acked-by: Namhyung Kim 
Cc: Athira Rajeev 
Cc: Hari Bathini 
Cc: Ian Rogers 
Cc: Jiri Olsa 
Cc: linuxppc-dev@lists.ozlabs.org
Cc: Madhavan Srinivasan 
Cc: Michael Petlan 
Cc: Shivani Nittor 
Cc: Tanushree Shah 
Cc: Tejas Manhas 
Cc: Thomas Richter 
Signed-off-by: Arnaldo Carvalho de Melo

perf unwind-libunwind: Make libunwind register reading cross platform

2026-05-15T19:18:36+00:00

Move the libunwind register to perf register mapping functions in
arch/../util/unwind-libunwind.c into a new libunwind-arch
directory. Rename the functions to
__get_perf_regnum_for_unw_regnum_. Add untested ppc32 and s390
functions. Add a get_perf_regnum_for_unw_regnum function that takes an
ELF machine as well as a register number and chooses the appropriate
architecture implementation.

Split the x86 and powerpc 32 and 64-bit implementations apart so that
a single libunwind-.h header is included.

Move the e_machine into the unwind_info struct to make it easier to
pass.

Signed-off-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: Albert Ou 
Cc: Alexander Shishkin 
Cc: Alexandre Ghiti 
Cc: Andrew Jones 
Cc: Athira Rajeev 
Cc: Dapeng Mi 
Cc: Dmitrii Dolgov <9erthalion6@gmail.com>
Cc: Florian Fainelli 
Cc: Howard Chu 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Leo Yan 
Cc: Li Guan 
Cc: Namhyung Kim 
Cc: Palmer Dabbelt 
Cc: Paul Walmsley 
Cc: Peter Zijlstra 
Cc: Shimin Guo 
Cc: Thomas Richter 
Cc: Tomas Glozar 
Cc: Will Deacon 
[ Map UNW_PPC32_NIP to PERF_REG_POWERPC_NIP like done for 64-bit, pointed out by a local sashiko ]
Signed-off-by: Arnaldo Carvalho de Melo

perf tools: Use calloc() where applicable

2026-04-09T02:21:05+00:00

Instead of using zalloc(nr_entries * sizeof_entry) that is what calloc()
does.

In some places where linux/zalloc.h isn't needed, remove it, add when
needed and was getting it indirectly.

Signed-off-by: Arnaldo Carvalho de Melo 
Signed-off-by: Namhyung Kim

tools headers: Update the syscall tables and unistd.h, to support the new 'rseq_slice_yield' syscall

2026-03-05T20:20:23+00:00

Picking up the changes from these csets:

  2153b2e8917b73e9 ("sparc: Add architecture support for clone3")
  99d2592023e5d0a3 ("rseq: Implement sys_rseq_slice_yield()")
  4ac286c4a8d904c8 ("s390/syscalls: Switch to generic system call table generation")

This makes 'perf trace' support it, now its possible, for instance, to
do:

  # perf trace -e rseq_slice_yield --max-stack=16

Here is an example with the 'sendmmsg' syscall:

  root@x1:~# perf trace -e sendmmsg --max-stack 16 --max-events=1
       0.000 ( 0.062 ms): dbus-broker/1012 sendmmsg(fd: 150, mmsg: 0x7ffef57cca50, vlen: 1, flags: DONTWAIT|NOSIGNAL) = 1
                                         syscall_exit_to_user_mode_prepare ([kernel.kallsyms])
                                         syscall_exit_to_user_mode_prepare ([kernel.kallsyms])
                                         syscall_exit_to_user_mode ([kernel.kallsyms])
                                         do_syscall_64 ([kernel.kallsyms])
                                         entry_SYSCALL_64 ([kernel.kallsyms])
                                         [0x117ce7] (/usr/lib64/libc.so.6 (deleted))
  root@x1:~#

To do a system wide tracing of the new 'rseq_slice_yield' syscall with a
backtrace of at most 16 entries.

This addresses these perf tools build warnings:

  Warning: Kernel ABI header differences:
    diff -u tools/include/uapi/asm-generic/unistd.h include/uapi/asm-generic/unistd.h
    diff -u tools/scripts/syscall.tbl scripts/syscall.tbl
    diff -u tools/perf/arch/x86/entry/syscalls/syscall_32.tbl arch/x86/entry/syscalls/syscall_32.tbl
    diff -u tools/perf/arch/x86/entry/syscalls/syscall_64.tbl arch/x86/entry/syscalls/syscall_64.tbl
    diff -u tools/perf/arch/powerpc/entry/syscalls/syscall.tbl arch/powerpc/kernel/syscalls/syscall.tbl
    diff -u tools/perf/arch/s390/entry/syscalls/syscall.tbl arch/s390/kernel/syscalls/syscall.tbl
    diff -u tools/perf/arch/mips/entry/syscalls/syscall_n64.tbl arch/mips/kernel/syscalls/syscall_n64.tbl
    diff -u tools/perf/arch/arm/entry/syscalls/syscall.tbl arch/arm/tools/syscall.tbl
    diff -u tools/perf/arch/sh/entry/syscalls/syscall.tbl arch/sh/kernel/syscalls/syscall.tbl
    diff -u tools/perf/arch/sparc/entry/syscalls/syscall.tbl arch/sparc/kernel/syscalls/syscall.tbl
    diff -u tools/perf/arch/xtensa/entry/syscalls/syscall.tbl arch/xtensa/kernel/syscalls/syscall.tbl

Cc: Andreas Larsson 
Cc: Heiko Carstens 
Cc: Ludwig Rydberg 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Signed-off-by: Arnaldo Carvalho de Melo

perf regs: Remove __weak attributive arch_sdt_arg_parse_op() function

2026-02-06T15:16:12+00:00

In line with the previous patch, the __weak arch_sdt_arg_parse_op()
function is removed.

Architectural-specific implementations in the arch/ directory are now
converted into sub-functions within the util/perf-regs-arch/ directory.

The perf_sdt_arg_parse_op() function will call these sub-functions based
on the EM_HOST.

This change enables cross-architecture calls to arch_sdt_arg_parse_op().

No functional changes are intended.

Suggested-by: Ian Rogers 
Reviewed-by: Ian Rogers 
Signed-off-by: Dapeng Mi 
Cc: Adrian Hunter 
Cc: Albert Ou 
Cc: Alexander Shishkin 
Cc: Alexandre Ghiti 
Cc: Guo Ren 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: John Garry 
Cc: Mike Leach 
Cc: Namhyung Kim 
Cc: Palmer Dabbelt 
Cc: Paul Walmsley 
Cc: Peter Zijlstra 
Cc: Thomas Falcon 
Cc: Will Deacon 
Cc: Xudong Hao 
Cc: Zide Chen 
[ Fixed up somme fuzz with powerpc and x86 Build files wrt removing perf_regs.o ]
Signed-off-by: Arnaldo Carvalho de Melo

perf regs: Remove weak attributive archxxx_reg_mask() functions

2026-02-06T15:16:09+00:00

Currently, some architecture-specific perf-regs functions, such as
arch__intr_reg_mask() and arch__user_reg_mask(), are defined with the
__weak attribute.

This approach ensures that only functions matching the architecture of
the build/run host are compiled and executed, reducing build time and
binary size.

However, this __weak attribute restricts these functions to be called
only on the same architecture, preventing cross-architecture
functionality.

For example, a perf.data file captured on x86 cannot be parsed on an ARM
platform.

To address this limitation, this patch removes the __weak attribute from
these perf-regs functions.

The architecture-specific code is moved from the arch/ directory to the
util/perf-regs-arch/ directory.

The appropriate architectural functions are then called based on the
EM_HOST.

No functional changes are intended.

Suggested-by: Ian Rogers 
Reviewed-by: Ian Rogers 
Signed-off-by: Dapeng Mi 
Cc: Adrian Hunter 
Cc: Albert Ou 
Cc: Alexander Shishkin 
Cc: Alexandre Ghiti 
Cc: Guo Ren 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: John Garry 
Cc: Mike Leach 
Cc: Namhyung Kim 
Cc: Palmer Dabbelt 
Cc: Paul Walmsley 
Cc: Peter Zijlstra 
Cc: Thomas Falcon 
Cc: Will Deacon 
Cc: Xudong Hao 
Cc: Zide Chen 
[ Fixed up somme fuzz with s390 and riscv Build files wrt removing perf_regs.o ]
Signed-off-by: Arnaldo Carvalho de Melo

perf arch: Update arch headers to use relative UAPI paths

2026-02-06T15:15:48+00:00

The architectural specific headers perf_regs.h currently rely on the
host architecture's 'asm/perf_regs.h'.

This can lead to compilation inconsistencies or failures when including
and building perf for a target architecture that differs from the host's
architecture.

Explicitly point to the UAPI headers within the tools source tree using
relative paths.

This ensures that perf is always built against the intended
architecture.

No functional changes are intended.

Reviewed-by: Ian Rogers 
Signed-off-by: Dapeng Mi 
Cc: Adrian Hunter 
Cc: Albert Ou 
Cc: Alexander Shishkin 
Cc: Alexandre Ghiti 
Cc: Guo Ren 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: John Garry 
Cc: Mike Leach 
Cc: Namhyung Kim 
Cc: Palmer Dabbelt 
Cc: Paul Walmsley 
Cc: Peter Zijlstra 
Cc: Thomas Falcon 
Cc: Will Deacon 
Cc: Xudong Hao 
Cc: Zide Chen 
Signed-off-by: Arnaldo Carvalho de Melo

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 regs: Refactor use of arch__sample_reg_masks() to perf_reg_name()

2026-01-23T19:58:39+00:00

arch__sample_reg_masks isn't supported on ARM(32), csky, loongarch,
MIPS, RISC-V and s390.

The table returned by the function just has the name of a register
paired with the corresponding sample_regs_user mask value.

For a given perf register we can compute the name with perf_reg_name and
the mask is just 1 left-shifted by the perf register number.

Change __parse_regs to use this method for finding registers rather than
arch__sample_reg_masks, thereby adding __parse_regs support for ARM(32),
csky, loongarch, MIPS, RISC-V and s390.

As arch__sample_reg_masks is then unused, remove the now unneeded
declarations.

Signed-off-by: Ian Rogers 
Tested-by: Thomas Richter 
Cc: Aditya Bodkhe 
Cc: Adrian Hunter 
Cc: Albert Ou 
Cc: Alexandre Ghiti 
Cc: Andi Kleen 
Cc: Athira Rajeev 
Cc: Chun-Tse Shao 
Cc: Dapeng Mi 
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: 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