Switch 3 more cases where the perf_env arch was being used to
determine the architecture type and switch to using the ELF
machine. This includes the addition of minimal instruction parsing in
perf script for architectures other than x86, or x86 when
cross-platform.

Note: this series will conflict with the e_flags series:
https://lore.kernel.org/lkml/20260123222209.1181249-1-irogers@google.com/
that adds an additional out argument to thread__e_machine. In all the
cases here that argument should be NULL. So:
  thread__e_machine(thread, machine)
becomes:
  thread__e_machine(thread, machine, /*e_flags=*/NULL)

Ian Rogers (3):
  perf script: Fix script_fetch_insn for more than just x86
  perf callchain: Switch callchain_param_setup from an arch to an
    e_machine
  perf thread-stack: Switch thread_stack__init to use e_machine

 tools/perf/arch/x86/util/Build                |  1 -
 tools/perf/arch/x86/util/archinsn.c           | 27 -------
 tools/perf/builtin-report.c                   |  6 +-
 tools/perf/builtin-script.c                   | 20 +-----
 .../scripts/python/Perf-Trace-Util/Context.c  |  2 +-
 tools/perf/tests/dlfilter-test.c              |  1 -
 tools/perf/util/archinsn.h                    | 12 ----
 tools/perf/util/callchain.c                   |  5 +-
 tools/perf/util/callchain.h                   |  2 +-
 tools/perf/util/dlfilter.c                    |  3 +-
 tools/perf/util/sample.c                      | 71 +++++++++++++++++++
 tools/perf/util/sample.h                      |  7 ++
 tools/perf/util/thread-stack.c                |  4 +-
 tools/perf/util/trace-event-scripting.c       | 16 -----
 tools/perf/util/trace-event.h                 |  3 -
 15 files changed, 91 insertions(+), 89 deletions(-)
 delete mode 100644 tools/perf/arch/x86/util/archinsn.c
 delete mode 100644 tools/perf/util/archinsn.h

-- 
2.52.0.457.g6b5491de43-goog

On Fri, Jan 23, 2026 at 04:16:08PM -0800, Ian Rogers wrote:
> Switch 3 more cases where the perf_env arch was being used to
> determine the architecture type and switch to using the ELF
> machine. This includes the addition of minimal instruction parsing in
> perf script for architectures other than x86, or x86 when
> cross-platform.
> 
> Note: this series will conflict with the e_flags series:
> https://lore.kernel.org/lkml/20260123222209.1181249-1-irogers@google.com/
> that adds an additional out argument to thread__e_machine. In all the
> cases here that argument should be NULL. So:
>   thread__e_machine(thread, machine)
> becomes:
>   thread__e_machine(thread, machine, /*e_flags=*/NULL)

This one isn't applying.

- Arnaldo
 
> Ian Rogers (3):
>   perf script: Fix script_fetch_insn for more than just x86
>   perf callchain: Switch callchain_param_setup from an arch to an
>     e_machine
>   perf thread-stack: Switch thread_stack__init to use e_machine
> 
>  tools/perf/arch/x86/util/Build                |  1 -
>  tools/perf/arch/x86/util/archinsn.c           | 27 -------
>  tools/perf/builtin-report.c                   |  6 +-
>  tools/perf/builtin-script.c                   | 20 +-----
>  .../scripts/python/Perf-Trace-Util/Context.c  |  2 +-
>  tools/perf/tests/dlfilter-test.c              |  1 -
>  tools/perf/util/archinsn.h                    | 12 ----
>  tools/perf/util/callchain.c                   |  5 +-
>  tools/perf/util/callchain.h                   |  2 +-
>  tools/perf/util/dlfilter.c                    |  3 +-
>  tools/perf/util/sample.c                      | 71 +++++++++++++++++++
>  tools/perf/util/sample.h                      |  7 ++
>  tools/perf/util/thread-stack.c                |  4 +-
>  tools/perf/util/trace-event-scripting.c       | 16 -----
>  tools/perf/util/trace-event.h                 |  3 -
>  15 files changed, 91 insertions(+), 89 deletions(-)
>  delete mode 100644 tools/perf/arch/x86/util/archinsn.c
>  delete mode 100644 tools/perf/util/archinsn.h
> 
> -- 
> 2.52.0.457.g6b5491de43-goog
>

On Mon, Jan 26, 2026 at 05:52:39PM -0300, Arnaldo Carvalho de Melo wrote:
> On Fri, Jan 23, 2026 at 04:16:08PM -0800, Ian Rogers wrote:
> > Switch 3 more cases where the perf_env arch was being used to
> > determine the architecture type and switch to using the ELF
> > machine. This includes the addition of minimal instruction parsing in
> > perf script for architectures other than x86, or x86 when
> > cross-platform.
> > 
> > Note: this series will conflict with the e_flags series:
> > https://lore.kernel.org/lkml/20260123222209.1181249-1-irogers@google.com/
> > that adds an additional out argument to thread__e_machine. In all the
> > cases here that argument should be NULL. So:
> >   thread__e_machine(thread, machine)
> > becomes:
> >   thread__e_machine(thread, machine, /*e_flags=*/NULL)
> 
> This one isn't applying.

⬢ [acme@toolbx perf-tools-next]$        git am ./20260123_irogers_perf_more_elf_machine_usage.mbx
Applying: perf script: Fix script_fetch_insn for more than just x86
Applying: perf callchain: Switch callchain_param_setup from an arch to an e_machine
error: patch failed: tools/perf/util/callchain.c:31
error: tools/perf/util/callchain.c: patch does not apply
Patch failed at 0002 perf callchain: Switch callchain_param_setup from an arch to an e_machine
hint: Use 'git am --show-current-patch=diff' to see the failed patch
hint: When you have resolved this problem, run "git am --continue".
hint: If you prefer to skip this patch, run "git am --skip" instead.
hint: To restore the original branch and stop patching, run "git am --abort".
hint: Disable this message with "git config set advice.mergeConflict false"
⬢ [acme@toolbx perf-tools-next]$

- Arnaldo

On Mon, Jan 26, 2026 at 05:53:09PM -0300, Arnaldo Carvalho de Melo wrote:
> On Mon, Jan 26, 2026 at 05:52:39PM -0300, Arnaldo Carvalho de Melo wrote:
> > On Fri, Jan 23, 2026 at 04:16:08PM -0800, Ian Rogers wrote:
> > > Switch 3 more cases where the perf_env arch was being used to
> > > determine the architecture type and switch to using the ELF
> > > machine. This includes the addition of minimal instruction parsing in
> > > perf script for architectures other than x86, or x86 when
> > > cross-platform.
> > > 
> > > Note: this series will conflict with the e_flags series:
> > > https://lore.kernel.org/lkml/20260123222209.1181249-1-irogers@google.com/
> > > that adds an additional out argument to thread__e_machine. In all the
> > > cases here that argument should be NULL. So:
> > >   thread__e_machine(thread, machine)
> > > becomes:
> > >   thread__e_machine(thread, machine, /*e_flags=*/NULL)
> > 
> > This one isn't applying.
> 
> ⬢ [acme@toolbx perf-tools-next]$        git am ./20260123_irogers_perf_more_elf_machine_usage.mbx
> Applying: perf script: Fix script_fetch_insn for more than just x86
> Applying: perf callchain: Switch callchain_param_setup from an arch to an e_machine

So, please try to refresh it with what is in tmp.perf-tools-next now.

⬢ [acme@toolbx perf-tools-next]$ git log --oneline -10
0a6fb6604746c92b (HEAD -> perf-tools-next, x1/perf-tools-next, x1/HEAD, perf-tools-next/tmp.perf-tools-next, five/perf-tools-next, five/HEAD, acme/tmp.perf-tools-next) perf session: Print all machines in session dump
2becdd163ab37c9d perf unwind-libdw: Wire up e_flags for CSKY
0403930f7b1534e3 perf perf_regs: Accurately compute register names for CSKY
4e66527f8859a661 perf thread: Add optional e_flags output argument to thread__e_machine
7d0ebeb6c0f735d4 perf dso: Factor out e_machine reading for use in thread
76b2cf07a6d2a836 perf vendor events amd: Fix Zen 5 MAB allocation events
008603bda19b2968 perf test: Fix test perf evlist for z/VM s390x
dda5f926a1006c73 perf annotate: Fix BUILD_NONDISTRO=1 missing args->ms conversions to pointer
9966b382d06733f7 tools headers: Don't check arm64's unistd.h
129bb23a6f7d0226 Revert "perf tools: Fix arm64 build by generating unistd_64.h"
⬢ [acme@toolbx perf-tools-next]$

- Arnaldo

On Mon, Jan 26, 2026 at 1:35 PM Arnaldo Carvalho de Melo
<acme@kernel.org> wrote:
>
> On Mon, Jan 26, 2026 at 05:53:09PM -0300, Arnaldo Carvalho de Melo wrote:
> > On Mon, Jan 26, 2026 at 05:52:39PM -0300, Arnaldo Carvalho de Melo wrote:
> > > On Fri, Jan 23, 2026 at 04:16:08PM -0800, Ian Rogers wrote:
> > > > Switch 3 more cases where the perf_env arch was being used to
> > > > determine the architecture type and switch to using the ELF
> > > > machine. This includes the addition of minimal instruction parsing in
> > > > perf script for architectures other than x86, or x86 when
> > > > cross-platform.
> > > >
> > > > Note: this series will conflict with the e_flags series:
> > > > https://lore.kernel.org/lkml/20260123222209.1181249-1-irogers@google.com/
> > > > that adds an additional out argument to thread__e_machine. In all the
> > > > cases here that argument should be NULL. So:
> > > >   thread__e_machine(thread, machine)
> > > > becomes:
> > > >   thread__e_machine(thread, machine, /*e_flags=*/NULL)
> > >
> > > This one isn't applying.
> >
> > ⬢ [acme@toolbx perf-tools-next]$        git am ./20260123_irogers_perf_more_elf_machine_usage.mbx
> > Applying: perf script: Fix script_fetch_insn for more than just x86
> > Applying: perf callchain: Switch callchain_param_setup from an arch to an e_machine
>
> So, please try to refresh it with what is in tmp.perf-tools-next now.
>
> ⬢ [acme@toolbx perf-tools-next]$ git log --oneline -10
> 0a6fb6604746c92b (HEAD -> perf-tools-next, x1/perf-tools-next, x1/HEAD, perf-tools-next/tmp.perf-tools-next, five/perf-tools-next, five/HEAD, acme/tmp.perf-tools-next) perf session: Print all machines in session dump
> 2becdd163ab37c9d perf unwind-libdw: Wire up e_flags for CSKY
> 0403930f7b1534e3 perf perf_regs: Accurately compute register names for CSKY
> 4e66527f8859a661 perf thread: Add optional e_flags output argument to thread__e_machine
> 7d0ebeb6c0f735d4 perf dso: Factor out e_machine reading for use in thread
> 76b2cf07a6d2a836 perf vendor events amd: Fix Zen 5 MAB allocation events
> 008603bda19b2968 perf test: Fix test perf evlist for z/VM s390x
> dda5f926a1006c73 perf annotate: Fix BUILD_NONDISTRO=1 missing args->ms conversions to pointer
> 9966b382d06733f7 tools headers: Don't check arm64's unistd.h
> 129bb23a6f7d0226 Revert "perf tools: Fix arm64 build by generating unistd_64.h"
> ⬢ [acme@toolbx perf-tools-next]$

Thanks, sent now:
https://lore.kernel.org/lkml/20260126220550.2573257-1-irogers@google.com/

Ian

> - Arnaldo