Re: [PATCH] perf evsel: Improve callchain warning for s390

[Ian Rogers]

On Tue, Sep 22, 2026 at 7:43 AM Jens Remus <jremus@xxxxxxxxxxxxx> wrote:
>
> On s390 the kernel uses s390 back chain instead of frame pointers for
> stack tracing of user space since v6.7 commit aa44433ac4ee ("s390: add
> USER_STACKTRACE support").  This is because frame pointers on s390
> cannot be used for stack tracing. [1]
>
> This requires user space to maintain a s390 back chain.  For instance
> user space to be built with compiler option '-mbackchain' (instead of
> '-fno-omit-frame-pointer' used on other architectures, which should
> better not be used on s390 [1]).  Only few distributions and users build
> user space with '-mbackchain'.  Therefore '--call-graph fp' may in
> practice not produce the expected results.
>
> Commit ca76fb67ebdd ("perf evlist: Improve default event for s390")
> changed the default '-g' option to 'dwarf' on s390 and added a warning
> for s390 that wrongly claimed that "Framepointer unwinding lacks kernel
> support".  This warning resulted from a misinterpretation of comments
> in s390 cpumsf_pmu_event_init().
>
> The restriction in cpumsf_pmu_event_init() applies to callchain sampling
> with the s390 CPU Measurement Sampling Facility (CPUMSF) hardware PMU.
> CPUMSF events, such as 'cycles', do not support callchain sampling.
> This is independent of whether perf uses the 'fp' or 'dwarf' call-graph
> mode.  The CPUMSF PMU provides samples collected asynchronously, making
> it impossible to associate a callchain to the historic IPs.  Therefore
> callchains can only be used with software events on s390.
>
> Remove the incorrect warning.  The kernel supports the 'fp' call-graph
> mode on s390 by walking the back chain.  Do not replace it with a
> hint to use 'dwarf' when the resulting callchain is incomplete.  Such a
> suggestion could imply that 'fp' is inherently inferior.  The same
> general limitation exists on other architectures when user space is not
> built with frame pointers (i.e. '-fno-omit-frame-pointers').
>
> Instead, warn on s390 when callchain sampling is requested with an event
> provided by the CPUMCF or CPUMSF hardware PMUs, because that combination
> is not supported.
>
> Update the perf record documentation to state that 'dwarf' is the
> default call-graph mode on s390 and that 'fp' uses back chain instead
> of frame pointers on s390.
>
> Note that '--call-graph fp' may also be useful for other applications,
> such as OpenJDK maintaining a s390 back chain (does not require JVM
> option '-XX:+PreserveFramePointer' on s390):
>
>   $ perf record --call-graph fp ... -- \
>     java -XX:+UnlockDiagnosticVMOptions -XX:+DumpPerfMapAtExit ...
>
> [1]: s390: Stack tracing using Frame Pointer, Back Chain, and SFrame,
>      https://conf.gnu-tools-cauldron.org/opo25/talk/Y3CVHY/
>
> Fixes: ca76fb67ebdd ("perf evlist: Improve default event for s390")
> Signed-off-by: Jens Remus <jremus@xxxxxxxxxxxxx>

Acked-by: Ian Rogers <irogers@xxxxxxxxxx>

Thanks,
Ian

> ---
>
> Notes (jremus):
>     Changes in v3:
>     - Updated warning message.
>     - Rebased on v7.3-rc4.
>
>  tools/perf/Documentation/perf-record.txt | 7 ++++++-
>  tools/perf/util/evsel.c                  | 6 +++---
>  2 files changed, 9 insertions(+), 4 deletions(-)
>
> diff --git a/tools/perf/Documentation/perf-record.txt b/tools/perf/Documentation/perf-record.txt
> index 178f483140ed..dc72606cb25d 100644
> --- a/tools/perf/Documentation/perf-record.txt
> +++ b/tools/perf/Documentation/perf-record.txt
> @@ -293,7 +293,7 @@ OPTIONS
>
>  --call-graph::
>         Setup and enable call-graph (stack chain/backtrace) recording,
> -       implies -g.  Default is "fp" (for user space).
> +       implies -g.  Default is "fp" for user space (s390 uses "dwarf").
>
>         The unwinding method used for kernel space is dependent on the
>         unwinder used by the active kernel configuration, i.e
> @@ -329,6 +329,11 @@ OPTIONS
>         enable deferred user callchain which will collect user-space callchains
>         when the thread returns to the user space.
>
> +       On s390, the "fp" method uses back chain instead of frame pointers,
> +       which requires user space to maintain a back chain (e.g. compiler
> +       option -mbackchain).  Without a maintained back chain this produces
> +       bogus call graphs.  In that case use "dwarf" instead.
> +
>  -q::
>  --quiet::
>         Don't print any warnings or messages, useful for scripting.
> diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
> index d4cb455f4a7d..55297b8dab37 100644
> --- a/tools/perf/util/evsel.c
> +++ b/tools/perf/util/evsel.c
> @@ -1092,9 +1092,9 @@ static void __evsel__config_callchain(struct evsel *evsel, const struct record_o
>         bool function = evsel__is_function_event(evsel);
>         struct perf_event_attr *attr = &evsel->core.attr;
>
> -       if (EM_HOST == EM_S390 && param->record_mode == CALLCHAIN_FP) {
> -               pr_warning_once(
> -                       "Framepointer unwinding lacks kernel support. Use '--call-graph dwarf'\n");
> +       if (EM_HOST == EM_S390 && evsel->pmu && !perf_pmu__is_software(evsel->pmu)) {
> +               pr_warning_once("Cannot use hardware PMU 'cpum_cf' / 'cpum_sf' event with callchain. "
> +                               "Use a software 'cpu-clock' / 'task-clock' event.\n");
>         }
>
>         evsel__set_sample_bit(evsel, CALLCHAIN);
>
> base-commit: 93f51579e7df248780214094418f205253383cc5
> --
> 2.53.0
>