tools/perf/builtin-annotate.c | 28 +- tools/perf/builtin-c2c.c | 6 +- tools/perf/builtin-diff.c | 5 +- tools/perf/builtin-inject.c | 75 +++--- tools/perf/builtin-kmem.c | 95 ++++--- tools/perf/builtin-kvm.c | 22 +- tools/perf/builtin-kwork.c | 253 +++++++++++------- tools/perf/builtin-lock.c | 128 +++++---- tools/perf/builtin-mem.c | 1 - tools/perf/builtin-record.c | 3 +- tools/perf/builtin-report.c | 38 ++- tools/perf/builtin-sched.c | 198 +++++++------- tools/perf/builtin-script.c | 22 +- tools/perf/builtin-timechart.c | 170 +++++++----- tools/perf/builtin-top.c | 19 +- tools/perf/builtin-trace.c | 121 +++++---- tools/perf/tests/hists_cumulate.c | 1 - tools/perf/tests/hists_filter.c | 1 - tools/perf/tests/hists_output.c | 1 - tools/perf/tests/mmap-basic.c | 4 +- tools/perf/tests/openat-syscall-tp-fields.c | 2 +- tools/perf/tests/switch-tracking.c | 9 +- tools/perf/util/annotate.c | 19 +- tools/perf/util/annotate.h | 6 +- tools/perf/util/bpf_kwork.c | 14 +- tools/perf/util/build-id.c | 3 +- tools/perf/util/build-id.h | 7 +- tools/perf/util/callchain.c | 8 +- tools/perf/util/callchain.h | 6 +- tools/perf/util/data-convert-bt.c | 2 +- tools/perf/util/data-convert-json.c | 5 +- tools/perf/util/db-export.c | 13 +- tools/perf/util/db-export.h | 3 +- tools/perf/util/evsel.c | 137 +++++++--- tools/perf/util/evsel.h | 12 +- tools/perf/util/hist.c | 26 +- tools/perf/util/hist.h | 3 +- tools/perf/util/intel-pt.c | 2 +- tools/perf/util/intel-tpebs.c | 3 +- tools/perf/util/jitdump.c | 2 +- .../perf/util/kvm-stat-arch/kvm-stat-arm64.c | 19 +- .../util/kvm-stat-arch/kvm-stat-loongarch.c | 17 +- .../util/kvm-stat-arch/kvm-stat-powerpc.c | 17 +- .../perf/util/kvm-stat-arch/kvm-stat-riscv.c | 17 +- tools/perf/util/kvm-stat-arch/kvm-stat-s390.c | 20 +- tools/perf/util/kvm-stat-arch/kvm-stat-x86.c | 70 ++--- tools/perf/util/kvm-stat.c | 19 +- tools/perf/util/kvm-stat.h | 18 +- tools/perf/util/kwork.h | 11 +- tools/perf/util/machine.c | 14 +- tools/perf/util/machine.h | 3 - tools/perf/util/s390-sample-raw.c | 31 ++- .../util/scripting-engines/trace-event-perl.c | 23 +- .../scripting-engines/trace-event-python.c | 47 ++-- tools/perf/util/session.c | 38 +-- tools/perf/util/synthetic-events.c | 49 ++-- tools/perf/util/synthetic-events.h | 2 - tools/perf/util/tool.c | 4 +- tools/perf/util/tool.h | 4 +- tools/perf/util/trace-event-scripting.c | 5 +- tools/perf/util/trace-event.h | 3 - 61 files changed, 991 insertions(+), 913 deletions(-)
Nearly all perf code ends up passing an evsel with the perf_sample,
which is problematic if you want to rewrite the evsel such as with
off-CPU processing - all uses of the evsel need fixing up. Previously
I'd mailed this patch as an RFC with everything combined:
https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/
and there was a request to break it up.
v13: Rebase removing patches merged by Namhyung. Namhyung proposed
changes that are in this series, but we became diverged. Address
some minor Sashiko issues.
v12: Add missing byteswap for tracepoint rawptr for dynamic
offsets. Avoid some unsigned <0 compiler warnings.
https://lore.kernel.org/linux-perf-users/20260413041143.1736055-1-irogers@google.com/
v11: Separate any remaining error path or other fixes out of
refactoring the evsel into sample. Incorporate Sashiko
feedback. Improve bounds checking for tracepoint data in
samples. Incorporate changes from Namhyung.
https://lore.kernel.org/lkml/20260413012227.1089445-1-irogers@google.com/
v10: Changes to some of the bounds checks as prompted by Sashiko.
https://lore.kernel.org/linux-perf-users/20260412020833.641177-1-irogers@google.com/
v9: Clean up some error path nits caught by Sashiko and simple bound
checks Sashiko recommended.
https://lore.kernel.org/linux-perf-users/20260411191710.524998-1-irogers@google.com/
v8: Fix address sanitizer issues with kwork and move patches, along
with bound check additions, to end of series.
https://lore.kernel.org/linux-perf-users/20260411065718.372240-1-irogers@google.com/
v7: Separate out kwork work->name to a strdup into its own patch, now patch 4.
https://lore.kernel.org/linux-perf-users/20260408072601.210837-1-irogers@google.com/
v6: Address more Sashiko feedback.
https://lore.kernel.org/lkml/20260404034325.3172592-1-irogers@google.com/
v5: Address feedback from Sashiko and Namhyung.
https://lore.kernel.org/lkml/20260403204017.2919994-1-irogers@google.com/
v4: Fix more sashiko issues: bounds checks, memory safety, making
refactors better, inconsistent evsels, mmap2 buildid injection
choice of evsel, mismatched function arguments.
https://sashiko.dev/#/patchset/20260320080835.724836-1-irogers%40google.com
https://lore.kernel.org/lkml/20260320192627.368357-1-irogers@google.com/
v3: Fix various sashiko review comments particularly about unintended
behavior changes.
https://lore.kernel.org/lkml/20260320080835.724836-1-irogers@google.com/
v2: Add review feedback on the first 2 patches from Namhyung, fix a
missed evsel assignment running event2evsel in builtin-inject.
https://lore.kernel.org/lkml/20260319232334.287517-1-irogers@google.com/
v1: https://lore.kernel.org/lkml/20260209174032.4142096-1-irogers@google.com/
Ian Rogers (32):
perf tool: Remove evsel from tool APIs that pass the sample
perf kvm: Don't pass evsel with sample
perf evsel: Refactor evsel tracepoint sample accessors perf_sample
perf trace: Don't pass evsel with sample
perf callchain: Don't pass evsel and sample
perf lock: Only pass sample to handlers
perf hist: Remove evsel parameter from inc samples functions
perf db-export: Remove evsel from struct export_sample
perf hist: Remove evsel from struct hist_entry_iter
perf report: Directly use sample->evsel to avoid computing from
sample->id
perf annotate: Don't pass evsel to add_sample
perf inject: Don't pass evsel with sample
perf kmem: Don't pass evsel with sample
perf kwork: Don't pass evsel with sample
perf sched: Don't pass evsel with sample
perf timechart: Don't pass evsel with sample
perf trace: Don't pass evsel with sample
perf evlist: Try to avoid computing evsel from sample
perf script: Don't pass evsel with sample
perf s390-sample-raw: Don't pass evsel or its PMU with sample
perf evsel: Don't pass evsel with sample
perf lock: Constify trace_lock_handler variables
perf lock: Avoid segv if event is missing a callchain
perf timechart: Fix memory leaks
perf kmem: Fix memory leaks on error path and when skipping
perf synthetic-events: Bound check when synthesizing mmap2 and
build_id events
perf kmem: Add bounds checks to tracepoint read values
perf sched: Bounds check CPU in sched switch events
perf timechart: Bounds check CPU
perf evsel: Add bounds checking to trace point raw data accessors
perf kwork: Fix address sanitizer issues
perf kwork: Fix memory management of kwork_work
tools/perf/builtin-annotate.c | 28 +-
tools/perf/builtin-c2c.c | 6 +-
tools/perf/builtin-diff.c | 5 +-
tools/perf/builtin-inject.c | 75 +++---
tools/perf/builtin-kmem.c | 95 ++++---
tools/perf/builtin-kvm.c | 22 +-
tools/perf/builtin-kwork.c | 253 +++++++++++-------
tools/perf/builtin-lock.c | 128 +++++----
tools/perf/builtin-mem.c | 1 -
tools/perf/builtin-record.c | 3 +-
tools/perf/builtin-report.c | 38 ++-
tools/perf/builtin-sched.c | 198 +++++++-------
tools/perf/builtin-script.c | 22 +-
tools/perf/builtin-timechart.c | 170 +++++++-----
tools/perf/builtin-top.c | 19 +-
tools/perf/builtin-trace.c | 121 +++++----
tools/perf/tests/hists_cumulate.c | 1 -
tools/perf/tests/hists_filter.c | 1 -
tools/perf/tests/hists_output.c | 1 -
tools/perf/tests/mmap-basic.c | 4 +-
tools/perf/tests/openat-syscall-tp-fields.c | 2 +-
tools/perf/tests/switch-tracking.c | 9 +-
tools/perf/util/annotate.c | 19 +-
tools/perf/util/annotate.h | 6 +-
tools/perf/util/bpf_kwork.c | 14 +-
tools/perf/util/build-id.c | 3 +-
tools/perf/util/build-id.h | 7 +-
tools/perf/util/callchain.c | 8 +-
tools/perf/util/callchain.h | 6 +-
tools/perf/util/data-convert-bt.c | 2 +-
tools/perf/util/data-convert-json.c | 5 +-
tools/perf/util/db-export.c | 13 +-
tools/perf/util/db-export.h | 3 +-
tools/perf/util/evsel.c | 137 +++++++---
tools/perf/util/evsel.h | 12 +-
tools/perf/util/hist.c | 26 +-
tools/perf/util/hist.h | 3 +-
tools/perf/util/intel-pt.c | 2 +-
tools/perf/util/intel-tpebs.c | 3 +-
tools/perf/util/jitdump.c | 2 +-
.../perf/util/kvm-stat-arch/kvm-stat-arm64.c | 19 +-
.../util/kvm-stat-arch/kvm-stat-loongarch.c | 17 +-
.../util/kvm-stat-arch/kvm-stat-powerpc.c | 17 +-
.../perf/util/kvm-stat-arch/kvm-stat-riscv.c | 17 +-
tools/perf/util/kvm-stat-arch/kvm-stat-s390.c | 20 +-
tools/perf/util/kvm-stat-arch/kvm-stat-x86.c | 70 ++---
tools/perf/util/kvm-stat.c | 19 +-
tools/perf/util/kvm-stat.h | 18 +-
tools/perf/util/kwork.h | 11 +-
tools/perf/util/machine.c | 14 +-
tools/perf/util/machine.h | 3 -
tools/perf/util/s390-sample-raw.c | 31 ++-
.../util/scripting-engines/trace-event-perl.c | 23 +-
.../scripting-engines/trace-event-python.c | 47 ++--
tools/perf/util/session.c | 38 +--
tools/perf/util/synthetic-events.c | 49 ++--
tools/perf/util/synthetic-events.h | 2 -
tools/perf/util/tool.c | 4 +-
tools/perf/util/tool.h | 4 +-
tools/perf/util/trace-event-scripting.c | 5 +-
tools/perf/util/trace-event.h | 3 -
61 files changed, 991 insertions(+), 913 deletions(-)
--
2.54.0.563.g4f69b47b94-goog
On Tue, May 12, 2026 at 03:29:29PM -0700, Ian Rogers wrote: > Nearly all perf code ends up passing an evsel with the perf_sample, > which is problematic if you want to rewrite the evsel such as with > off-CPU processing - all uses of the evsel need fixing up. Previously > I'd mailed this patch as an RFC with everything combined: > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > and there was a request to break it up. > > v13: Rebase removing patches merged by Namhyung. Namhyung proposed > changes that are in this series, but we became diverged. Address > some minor Sashiko issues. > > v12: Add missing byteswap for tracepoint rawptr for dynamic > offsets. Avoid some unsigned <0 compiler warnings. > https://lore.kernel.org/linux-perf-users/20260413041143.1736055-1-irogers@google.com/ > > v11: Separate any remaining error path or other fixes out of > refactoring the evsel into sample. Incorporate Sashiko > feedback. Improve bounds checking for tracepoint data in > samples. Incorporate changes from Namhyung. > https://lore.kernel.org/lkml/20260413012227.1089445-1-irogers@google.com/ > > v10: Changes to some of the bounds checks as prompted by Sashiko. > https://lore.kernel.org/linux-perf-users/20260412020833.641177-1-irogers@google.com/ > > v9: Clean up some error path nits caught by Sashiko and simple bound > checks Sashiko recommended. > https://lore.kernel.org/linux-perf-users/20260411191710.524998-1-irogers@google.com/ > > v8: Fix address sanitizer issues with kwork and move patches, along > with bound check additions, to end of series. > https://lore.kernel.org/linux-perf-users/20260411065718.372240-1-irogers@google.com/ > > v7: Separate out kwork work->name to a strdup into its own patch, now patch 4. > https://lore.kernel.org/linux-perf-users/20260408072601.210837-1-irogers@google.com/ > > v6: Address more Sashiko feedback. > https://lore.kernel.org/lkml/20260404034325.3172592-1-irogers@google.com/ > > v5: Address feedback from Sashiko and Namhyung. > https://lore.kernel.org/lkml/20260403204017.2919994-1-irogers@google.com/ > > v4: Fix more sashiko issues: bounds checks, memory safety, making > refactors better, inconsistent evsels, mmap2 buildid injection > choice of evsel, mismatched function arguments. > https://sashiko.dev/#/patchset/20260320080835.724836-1-irogers%40google.com > https://lore.kernel.org/lkml/20260320192627.368357-1-irogers@google.com/ > > v3: Fix various sashiko review comments particularly about unintended > behavior changes. > https://lore.kernel.org/lkml/20260320080835.724836-1-irogers@google.com/ > > v2: Add review feedback on the first 2 patches from Namhyung, fix a > missed evsel assignment running event2evsel in builtin-inject. > https://lore.kernel.org/lkml/20260319232334.287517-1-irogers@google.com/ > > v1: https://lore.kernel.org/lkml/20260209174032.4142096-1-irogers@google.com/ > > Ian Rogers (32): > perf tool: Remove evsel from tool APIs that pass the sample > perf kvm: Don't pass evsel with sample > perf evsel: Refactor evsel tracepoint sample accessors perf_sample > perf trace: Don't pass evsel with sample > perf callchain: Don't pass evsel and sample > perf lock: Only pass sample to handlers > perf hist: Remove evsel parameter from inc samples functions > perf db-export: Remove evsel from struct export_sample > perf hist: Remove evsel from struct hist_entry_iter > perf report: Directly use sample->evsel to avoid computing from > sample->id > perf annotate: Don't pass evsel to add_sample > perf inject: Don't pass evsel with sample > perf kmem: Don't pass evsel with sample > perf kwork: Don't pass evsel with sample > perf sched: Don't pass evsel with sample > perf timechart: Don't pass evsel with sample > perf trace: Don't pass evsel with sample > perf evlist: Try to avoid computing evsel from sample > perf script: Don't pass evsel with sample > perf s390-sample-raw: Don't pass evsel or its PMU with sample > perf evsel: Don't pass evsel with sample > perf lock: Constify trace_lock_handler variables > perf lock: Avoid segv if event is missing a callchain > perf timechart: Fix memory leaks > perf kmem: Fix memory leaks on error path and when skipping > perf synthetic-events: Bound check when synthesizing mmap2 and > build_id events > perf kmem: Add bounds checks to tracepoint read values > perf sched: Bounds check CPU in sched switch events > perf timechart: Bounds check CPU > perf evsel: Add bounds checking to trace point raw data accessors > perf kwork: Fix address sanitizer issues > perf kwork: Fix memory management of kwork_work I have a few nitpicks but I think it's good to go. Acked-by: Namhyung Kim <namhyung@kernel.org> Thanks, Namhyung > > tools/perf/builtin-annotate.c | 28 +- > tools/perf/builtin-c2c.c | 6 +- > tools/perf/builtin-diff.c | 5 +- > tools/perf/builtin-inject.c | 75 +++--- > tools/perf/builtin-kmem.c | 95 ++++--- > tools/perf/builtin-kvm.c | 22 +- > tools/perf/builtin-kwork.c | 253 +++++++++++------- > tools/perf/builtin-lock.c | 128 +++++---- > tools/perf/builtin-mem.c | 1 - > tools/perf/builtin-record.c | 3 +- > tools/perf/builtin-report.c | 38 ++- > tools/perf/builtin-sched.c | 198 +++++++------- > tools/perf/builtin-script.c | 22 +- > tools/perf/builtin-timechart.c | 170 +++++++----- > tools/perf/builtin-top.c | 19 +- > tools/perf/builtin-trace.c | 121 +++++---- > tools/perf/tests/hists_cumulate.c | 1 - > tools/perf/tests/hists_filter.c | 1 - > tools/perf/tests/hists_output.c | 1 - > tools/perf/tests/mmap-basic.c | 4 +- > tools/perf/tests/openat-syscall-tp-fields.c | 2 +- > tools/perf/tests/switch-tracking.c | 9 +- > tools/perf/util/annotate.c | 19 +- > tools/perf/util/annotate.h | 6 +- > tools/perf/util/bpf_kwork.c | 14 +- > tools/perf/util/build-id.c | 3 +- > tools/perf/util/build-id.h | 7 +- > tools/perf/util/callchain.c | 8 +- > tools/perf/util/callchain.h | 6 +- > tools/perf/util/data-convert-bt.c | 2 +- > tools/perf/util/data-convert-json.c | 5 +- > tools/perf/util/db-export.c | 13 +- > tools/perf/util/db-export.h | 3 +- > tools/perf/util/evsel.c | 137 +++++++--- > tools/perf/util/evsel.h | 12 +- > tools/perf/util/hist.c | 26 +- > tools/perf/util/hist.h | 3 +- > tools/perf/util/intel-pt.c | 2 +- > tools/perf/util/intel-tpebs.c | 3 +- > tools/perf/util/jitdump.c | 2 +- > .../perf/util/kvm-stat-arch/kvm-stat-arm64.c | 19 +- > .../util/kvm-stat-arch/kvm-stat-loongarch.c | 17 +- > .../util/kvm-stat-arch/kvm-stat-powerpc.c | 17 +- > .../perf/util/kvm-stat-arch/kvm-stat-riscv.c | 17 +- > tools/perf/util/kvm-stat-arch/kvm-stat-s390.c | 20 +- > tools/perf/util/kvm-stat-arch/kvm-stat-x86.c | 70 ++--- > tools/perf/util/kvm-stat.c | 19 +- > tools/perf/util/kvm-stat.h | 18 +- > tools/perf/util/kwork.h | 11 +- > tools/perf/util/machine.c | 14 +- > tools/perf/util/machine.h | 3 - > tools/perf/util/s390-sample-raw.c | 31 ++- > .../util/scripting-engines/trace-event-perl.c | 23 +- > .../scripting-engines/trace-event-python.c | 47 ++-- > tools/perf/util/session.c | 38 +-- > tools/perf/util/synthetic-events.c | 49 ++-- > tools/perf/util/synthetic-events.h | 2 - > tools/perf/util/tool.c | 4 +- > tools/perf/util/tool.h | 4 +- > tools/perf/util/trace-event-scripting.c | 5 +- > tools/perf/util/trace-event.h | 3 - > 61 files changed, 991 insertions(+), 913 deletions(-) > > -- > 2.54.0.563.g4f69b47b94-goog >
On Tue, May 19, 2026 at 2:34 PM Namhyung Kim <namhyung@kernel.org> wrote: > > On Tue, May 12, 2026 at 03:29:29PM -0700, Ian Rogers wrote: > > Nearly all perf code ends up passing an evsel with the perf_sample, > > which is problematic if you want to rewrite the evsel such as with > > off-CPU processing - all uses of the evsel need fixing up. Previously > > I'd mailed this patch as an RFC with everything combined: > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > > and there was a request to break it up. > > > > v13: Rebase removing patches merged by Namhyung. Namhyung proposed > > changes that are in this series, but we became diverged. Address > > some minor Sashiko issues. > > > > v12: Add missing byteswap for tracepoint rawptr for dynamic > > offsets. Avoid some unsigned <0 compiler warnings. > > https://lore.kernel.org/linux-perf-users/20260413041143.1736055-1-irogers@google.com/ > > > > v11: Separate any remaining error path or other fixes out of > > refactoring the evsel into sample. Incorporate Sashiko > > feedback. Improve bounds checking for tracepoint data in > > samples. Incorporate changes from Namhyung. > > https://lore.kernel.org/lkml/20260413012227.1089445-1-irogers@google.com/ > > > > v10: Changes to some of the bounds checks as prompted by Sashiko. > > https://lore.kernel.org/linux-perf-users/20260412020833.641177-1-irogers@google.com/ > > > > v9: Clean up some error path nits caught by Sashiko and simple bound > > checks Sashiko recommended. > > https://lore.kernel.org/linux-perf-users/20260411191710.524998-1-irogers@google.com/ > > > > v8: Fix address sanitizer issues with kwork and move patches, along > > with bound check additions, to end of series. > > https://lore.kernel.org/linux-perf-users/20260411065718.372240-1-irogers@google.com/ > > > > v7: Separate out kwork work->name to a strdup into its own patch, now patch 4. > > https://lore.kernel.org/linux-perf-users/20260408072601.210837-1-irogers@google.com/ > > > > v6: Address more Sashiko feedback. > > https://lore.kernel.org/lkml/20260404034325.3172592-1-irogers@google.com/ > > > > v5: Address feedback from Sashiko and Namhyung. > > https://lore.kernel.org/lkml/20260403204017.2919994-1-irogers@google.com/ > > > > v4: Fix more sashiko issues: bounds checks, memory safety, making > > refactors better, inconsistent evsels, mmap2 buildid injection > > choice of evsel, mismatched function arguments. > > https://sashiko.dev/#/patchset/20260320080835.724836-1-irogers%40google.com > > https://lore.kernel.org/lkml/20260320192627.368357-1-irogers@google.com/ > > > > v3: Fix various sashiko review comments particularly about unintended > > behavior changes. > > https://lore.kernel.org/lkml/20260320080835.724836-1-irogers@google.com/ > > > > v2: Add review feedback on the first 2 patches from Namhyung, fix a > > missed evsel assignment running event2evsel in builtin-inject. > > https://lore.kernel.org/lkml/20260319232334.287517-1-irogers@google.com/ > > > > v1: https://lore.kernel.org/lkml/20260209174032.4142096-1-irogers@google.com/ > > > > Ian Rogers (32): > > perf tool: Remove evsel from tool APIs that pass the sample > > perf kvm: Don't pass evsel with sample > > perf evsel: Refactor evsel tracepoint sample accessors perf_sample > > perf trace: Don't pass evsel with sample > > perf callchain: Don't pass evsel and sample > > perf lock: Only pass sample to handlers > > perf hist: Remove evsel parameter from inc samples functions > > perf db-export: Remove evsel from struct export_sample > > perf hist: Remove evsel from struct hist_entry_iter > > perf report: Directly use sample->evsel to avoid computing from > > sample->id > > perf annotate: Don't pass evsel to add_sample > > perf inject: Don't pass evsel with sample > > perf kmem: Don't pass evsel with sample > > perf kwork: Don't pass evsel with sample > > perf sched: Don't pass evsel with sample > > perf timechart: Don't pass evsel with sample > > perf trace: Don't pass evsel with sample > > perf evlist: Try to avoid computing evsel from sample > > perf script: Don't pass evsel with sample > > perf s390-sample-raw: Don't pass evsel or its PMU with sample > > perf evsel: Don't pass evsel with sample > > perf lock: Constify trace_lock_handler variables > > perf lock: Avoid segv if event is missing a callchain > > perf timechart: Fix memory leaks > > perf kmem: Fix memory leaks on error path and when skipping > > perf synthetic-events: Bound check when synthesizing mmap2 and > > build_id events > > perf kmem: Add bounds checks to tracepoint read values > > perf sched: Bounds check CPU in sched switch events > > perf timechart: Bounds check CPU > > perf evsel: Add bounds checking to trace point raw data accessors > > perf kwork: Fix address sanitizer issues > > perf kwork: Fix memory management of kwork_work > > I have a few nitpicks but I think it's good to go. > > Acked-by: Namhyung Kim <namhyung@kernel.org> Thanks Namhyung, I'll rebase, address nits, and add your tag. I'll keep an eye out in case Sashiko spots more issues, but we were pretty clean so hopefully those things can be follow-ups. Ian > Thanks, > Namhyung > > > > > tools/perf/builtin-annotate.c | 28 +- > > tools/perf/builtin-c2c.c | 6 +- > > tools/perf/builtin-diff.c | 5 +- > > tools/perf/builtin-inject.c | 75 +++--- > > tools/perf/builtin-kmem.c | 95 ++++--- > > tools/perf/builtin-kvm.c | 22 +- > > tools/perf/builtin-kwork.c | 253 +++++++++++------- > > tools/perf/builtin-lock.c | 128 +++++---- > > tools/perf/builtin-mem.c | 1 - > > tools/perf/builtin-record.c | 3 +- > > tools/perf/builtin-report.c | 38 ++- > > tools/perf/builtin-sched.c | 198 +++++++------- > > tools/perf/builtin-script.c | 22 +- > > tools/perf/builtin-timechart.c | 170 +++++++----- > > tools/perf/builtin-top.c | 19 +- > > tools/perf/builtin-trace.c | 121 +++++---- > > tools/perf/tests/hists_cumulate.c | 1 - > > tools/perf/tests/hists_filter.c | 1 - > > tools/perf/tests/hists_output.c | 1 - > > tools/perf/tests/mmap-basic.c | 4 +- > > tools/perf/tests/openat-syscall-tp-fields.c | 2 +- > > tools/perf/tests/switch-tracking.c | 9 +- > > tools/perf/util/annotate.c | 19 +- > > tools/perf/util/annotate.h | 6 +- > > tools/perf/util/bpf_kwork.c | 14 +- > > tools/perf/util/build-id.c | 3 +- > > tools/perf/util/build-id.h | 7 +- > > tools/perf/util/callchain.c | 8 +- > > tools/perf/util/callchain.h | 6 +- > > tools/perf/util/data-convert-bt.c | 2 +- > > tools/perf/util/data-convert-json.c | 5 +- > > tools/perf/util/db-export.c | 13 +- > > tools/perf/util/db-export.h | 3 +- > > tools/perf/util/evsel.c | 137 +++++++--- > > tools/perf/util/evsel.h | 12 +- > > tools/perf/util/hist.c | 26 +- > > tools/perf/util/hist.h | 3 +- > > tools/perf/util/intel-pt.c | 2 +- > > tools/perf/util/intel-tpebs.c | 3 +- > > tools/perf/util/jitdump.c | 2 +- > > .../perf/util/kvm-stat-arch/kvm-stat-arm64.c | 19 +- > > .../util/kvm-stat-arch/kvm-stat-loongarch.c | 17 +- > > .../util/kvm-stat-arch/kvm-stat-powerpc.c | 17 +- > > .../perf/util/kvm-stat-arch/kvm-stat-riscv.c | 17 +- > > tools/perf/util/kvm-stat-arch/kvm-stat-s390.c | 20 +- > > tools/perf/util/kvm-stat-arch/kvm-stat-x86.c | 70 ++--- > > tools/perf/util/kvm-stat.c | 19 +- > > tools/perf/util/kvm-stat.h | 18 +- > > tools/perf/util/kwork.h | 11 +- > > tools/perf/util/machine.c | 14 +- > > tools/perf/util/machine.h | 3 - > > tools/perf/util/s390-sample-raw.c | 31 ++- > > .../util/scripting-engines/trace-event-perl.c | 23 +- > > .../scripting-engines/trace-event-python.c | 47 ++-- > > tools/perf/util/session.c | 38 +-- > > tools/perf/util/synthetic-events.c | 49 ++-- > > tools/perf/util/synthetic-events.h | 2 - > > tools/perf/util/tool.c | 4 +- > > tools/perf/util/tool.h | 4 +- > > tools/perf/util/trace-event-scripting.c | 5 +- > > tools/perf/util/trace-event.h | 3 - > > 61 files changed, 991 insertions(+), 913 deletions(-) > > > > -- > > 2.54.0.563.g4f69b47b94-goog > >
Nearly all perf code ends up passing an evsel with the perf_sample,
which is problematic if you want to rewrite the evsel (such as with
off-CPU processing) because all uses of the evsel need to be fixed
up. Since the perf_sample now carries its own resolved evsel pointer,
passing evsel separately is redundant and error-prone (allowing
parameter divergence bugs).
This series cleans up the redundant evsel parameter across all perf
tool subcommands and APIs, ensuring that tools uniformly retrieve
event attributes directly from the sample itself. This simplifies
function signatures and improves API consistency.
Additionally, this series incorporates subsequent memory safety
improvements, robust bounds checking, and memory leak corrections
prompted by address sanitizers and Sashiko code reviews.
The original combined RFC patch:
https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/
was split up per reviewer requests.
v14:
- Incorporate Acked-by tags from Namhyung Kim.
- Re-polish and expand commit descriptions to provide thorough explanations
of the changes and their rationale.
v13: Rebase removing patches merged by Namhyung. Namhyung proposed
changes that are in this series, but we became diverged. Address
some minor Sashiko issues.
https://lore.kernel.org/linux-perf-users/20260512223001.2952848-1-irogers@google.com/
v12: Add missing byteswap for tracepoint rawptr for dynamic
offsets. Avoid some unsigned <0 compiler warnings.
https://lore.kernel.org/linux-perf-users/20260413041143.1736055-1-irogers@google.com/
v11: Separate any remaining error path or other fixes out of
refactoring the evsel into sample. Incorporate Sashiko
feedback. Improve bounds checking for tracepoint data in
samples. Incorporate changes from Namhyung.
https://lore.kernel.org/lkml/20260413012227.1089445-1-irogers@google.com/
v10: Changes to some of the bounds checks as prompted by Sashiko.
https://lore.kernel.org/linux-perf-users/20260412020833.641177-1-irogers@google.com/
v9: Clean up some error path nits caught by Sashiko and simple bound
checks Sashiko recommended.
https://lore.kernel.org/linux-perf-users/20260411191710.524998-1-irogers@google.com/
v8: Fix address sanitizer issues with kwork and move patches, along
with bound check additions, to end of series.
https://lore.kernel.org/linux-perf-users/20260411065718.372240-1-irogers@google.com/
v7: Separate out kwork work->name to a strdup into its own patch, now patch 4.
https://lore.kernel.org/linux-perf-users/20260408072601.210837-1-irogers@google.com/
v6: Address more Sashiko feedback.
https://lore.kernel.org/lkml/20260404034325.3172592-1-irogers@google.com/
v5: Address feedback from Sashiko and Namhyung.
https://lore.kernel.org/lkml/20260403204017.2919994-1-irogers@google.com/
v4: Fix more sashiko issues: bounds checks, memory safety, making
refactors better, inconsistent evsels, mmap2 buildid injection
choice of evsel, mismatched function arguments.
https://sashiko.dev/#/patchset/20260320080835.724836-1-irogers%40google.com
https://lore.kernel.org/lkml/20260320192627.368357-1-irogers@google.com/
v3: Fix various sashiko review comments particularly about unintended
behavior changes.
https://lore.kernel.org/lkml/20260320080835.724836-1-irogers@google.com/
v2: Add review feedback on the first 2 patches from Namhyung, fix a
missed evsel assignment running event2evsel in builtin-inject.
https://lore.kernel.org/lkml/20260319232334.287517-1-irogers@google.com/
v1: https://lore.kernel.org/lkml/20260209174032.4142096-1-irogers@google.com/
Ian Rogers (32):
perf tool: Remove evsel from tool APIs that pass the sample
perf kvm: Don't pass evsel with sample
perf evsel: Refactor evsel tracepoint sample accessors perf_sample
perf trace: Don't pass evsel with sample
perf callchain: Don't pass evsel and sample
perf lock: Only pass sample to handlers
perf hist: Remove evsel parameter from inc samples functions
perf db-export: Remove evsel from struct export_sample
perf hist: Remove evsel from struct hist_entry_iter
perf report: Directly use sample->evsel to avoid computing from
sample->id
perf annotate: Don't pass evsel to add_sample
perf inject: Don't pass evsel with sample
perf kmem: Don't pass evsel with sample
perf kwork: Don't pass evsel with sample
perf sched: Don't pass evsel with sample
perf timechart: Don't pass evsel with sample
perf trace: Don't pass evsel with sample
perf evlist: Try to avoid computing evsel from sample
perf script: Don't pass evsel with sample
perf s390-sample-raw: Don't pass evsel or its PMU with sample
perf evsel: Don't pass evsel with sample
perf lock: Constify trace_lock_handler variables
perf lock: Avoid segv if event is missing a callchain
perf timechart: Fix memory leaks
perf kmem: Fix memory leaks on error path and when skipping
perf synthetic-events: Bound check when synthesizing mmap2 and
build_id events
perf kmem: Add bounds checks to tracepoint read values
perf sched: Bounds check CPU in sched switch events
perf timechart: Bounds check CPU
perf evsel: Add bounds checking to trace point raw data accessors
perf kwork: Fix address sanitizer issues
perf kwork: Fix memory management of kwork_work
tools/perf/builtin-annotate.c | 28 +-
tools/perf/builtin-c2c.c | 6 +-
tools/perf/builtin-diff.c | 5 +-
tools/perf/builtin-inject.c | 75 +++---
tools/perf/builtin-kmem.c | 95 ++++---
tools/perf/builtin-kvm.c | 22 +-
tools/perf/builtin-kwork.c | 249 +++++++++++-------
tools/perf/builtin-lock.c | 128 +++++----
tools/perf/builtin-mem.c | 1 -
tools/perf/builtin-record.c | 3 +-
tools/perf/builtin-report.c | 38 ++-
tools/perf/builtin-sched.c | 198 +++++++-------
tools/perf/builtin-script.c | 22 +-
tools/perf/builtin-timechart.c | 164 +++++++-----
tools/perf/builtin-top.c | 19 +-
tools/perf/builtin-trace.c | 121 +++++----
tools/perf/tests/hists_cumulate.c | 1 -
tools/perf/tests/hists_filter.c | 1 -
tools/perf/tests/hists_output.c | 1 -
tools/perf/tests/mmap-basic.c | 4 +-
tools/perf/tests/openat-syscall-tp-fields.c | 2 +-
tools/perf/tests/switch-tracking.c | 9 +-
tools/perf/util/annotate.c | 19 +-
tools/perf/util/annotate.h | 6 +-
tools/perf/util/bpf_kwork.c | 22 +-
tools/perf/util/build-id.c | 3 +-
tools/perf/util/build-id.h | 7 +-
tools/perf/util/callchain.c | 8 +-
tools/perf/util/callchain.h | 6 +-
tools/perf/util/data-convert-bt.c | 2 +-
tools/perf/util/data-convert-json.c | 5 +-
tools/perf/util/db-export.c | 13 +-
tools/perf/util/db-export.h | 3 +-
tools/perf/util/evsel.c | 137 +++++++---
tools/perf/util/evsel.h | 12 +-
tools/perf/util/hist.c | 26 +-
tools/perf/util/hist.h | 3 +-
tools/perf/util/intel-pt.c | 2 +-
tools/perf/util/intel-tpebs.c | 3 +-
tools/perf/util/jitdump.c | 2 +-
.../perf/util/kvm-stat-arch/kvm-stat-arm64.c | 19 +-
.../util/kvm-stat-arch/kvm-stat-loongarch.c | 17 +-
.../util/kvm-stat-arch/kvm-stat-powerpc.c | 17 +-
.../perf/util/kvm-stat-arch/kvm-stat-riscv.c | 17 +-
tools/perf/util/kvm-stat-arch/kvm-stat-s390.c | 20 +-
tools/perf/util/kvm-stat-arch/kvm-stat-x86.c | 70 ++---
tools/perf/util/kvm-stat.c | 19 +-
tools/perf/util/kvm-stat.h | 18 +-
tools/perf/util/kwork.h | 11 +-
tools/perf/util/machine.c | 14 +-
tools/perf/util/machine.h | 3 -
tools/perf/util/s390-sample-raw.c | 31 ++-
.../util/scripting-engines/trace-event-perl.c | 23 +-
.../scripting-engines/trace-event-python.c | 47 ++--
tools/perf/util/session.c | 38 +--
tools/perf/util/synthetic-events.c | 49 ++--
tools/perf/util/synthetic-events.h | 2 -
tools/perf/util/tool.c | 4 +-
tools/perf/util/tool.h | 4 +-
tools/perf/util/trace-event-scripting.c | 5 +-
tools/perf/util/trace-event.h | 3 -
61 files changed, 991 insertions(+), 911 deletions(-)
--
2.54.0.746.g67dd491aae-goog
On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote: > Nearly all perf code ends up passing an evsel with the perf_sample, > which is problematic if you want to rewrite the evsel (such as with > off-CPU processing) because all uses of the evsel need to be fixed > up. Since the perf_sample now carries its own resolved evsel pointer, > passing evsel separately is redundant and error-prone (allowing > parameter divergence bugs). > > This series cleans up the redundant evsel parameter across all perf > tool subcommands and APIs, ensuring that tools uniformly retrieve > event attributes directly from the sample itself. This simplifies > function signatures and improves API consistency. > > Additionally, this series incorporates subsequent memory safety > improvements, robust bounds checking, and memory leak corrections > prompted by address sanitizers and Sashiko code reviews. > > The original combined RFC patch: > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > was split up per reviewer requests. > > v14: > - Incorporate Acked-by tags from Namhyung Kim. > - Re-polish and expand commit descriptions to provide thorough explanations > of the changes and their rationale. Thanks, tentatively applying it to perf-tools-next, for v7.2. Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
On Wed, May 20, 2026 at 04:15:30PM -0300, Arnaldo Carvalho de Melo wrote: > On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote: > > Nearly all perf code ends up passing an evsel with the perf_sample, > > which is problematic if you want to rewrite the evsel (such as with > > off-CPU processing) because all uses of the evsel need to be fixed > > up. Since the perf_sample now carries its own resolved evsel pointer, > > passing evsel separately is redundant and error-prone (allowing > > parameter divergence bugs). > > > > This series cleans up the redundant evsel parameter across all perf > > tool subcommands and APIs, ensuring that tools uniformly retrieve > > event attributes directly from the sample itself. This simplifies > > function signatures and improves API consistency. > > > > Additionally, this series incorporates subsequent memory safety > > improvements, robust bounds checking, and memory leak corrections > > prompted by address sanitizers and Sashiko code reviews. > > > > The original combined RFC patch: > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > > was split up per reviewer requests. > > > > v14: > > - Incorporate Acked-by tags from Namhyung Kim. > > - Re-polish and expand commit descriptions to provide thorough explanations > > of the changes and their rationale. > > Thanks, tentatively applying it to perf-tools-next, for v7.2. Had to apply a minor fixup for a conflict with anoter patch series from you, the LBR one. - Arnaldo
On Wed, May 20, 2026 at 12:46 PM Arnaldo Carvalho de Melo <acme@kernel.org> wrote: > > On Wed, May 20, 2026 at 04:15:30PM -0300, Arnaldo Carvalho de Melo wrote: > > On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote: > > > Nearly all perf code ends up passing an evsel with the perf_sample, > > > which is problematic if you want to rewrite the evsel (such as with > > > off-CPU processing) because all uses of the evsel need to be fixed > > > up. Since the perf_sample now carries its own resolved evsel pointer, > > > passing evsel separately is redundant and error-prone (allowing > > > parameter divergence bugs). > > > > > > This series cleans up the redundant evsel parameter across all perf > > > tool subcommands and APIs, ensuring that tools uniformly retrieve > > > event attributes directly from the sample itself. This simplifies > > > function signatures and improves API consistency. > > > > > > Additionally, this series incorporates subsequent memory safety > > > improvements, robust bounds checking, and memory leak corrections > > > prompted by address sanitizers and Sashiko code reviews. > > > > > > The original combined RFC patch: > > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > > > was split up per reviewer requests. > > > > > > v14: > > > - Incorporate Acked-by tags from Namhyung Kim. > > > - Re-polish and expand commit descriptions to provide thorough explanations > > > of the changes and their rationale. > > > > Thanks, tentatively applying it to perf-tools-next, for v7.2. > > Had to apply a minor fixup for a conflict with anoter patch series from > you, the LBR one. Sorry for that and thanks for dealing with the conflict! Any chance you could push to tmp.perf-tools-next? I have some other series that will conflict and it would be nice to rebase those changes there. Thanks, Ian > - Arnaldo
On Wed, May 20, 2026 at 01:40:48PM -0700, Ian Rogers wrote: > On Wed, May 20, 2026 at 12:46 PM Arnaldo Carvalho de Melo > <acme@kernel.org> wrote: > > > > On Wed, May 20, 2026 at 04:15:30PM -0300, Arnaldo Carvalho de Melo wrote: > > > On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote: > > > > Nearly all perf code ends up passing an evsel with the perf_sample, > > > > which is problematic if you want to rewrite the evsel (such as with > > > > off-CPU processing) because all uses of the evsel need to be fixed > > > > up. Since the perf_sample now carries its own resolved evsel pointer, > > > > passing evsel separately is redundant and error-prone (allowing > > > > parameter divergence bugs). > > > > > > > > This series cleans up the redundant evsel parameter across all perf > > > > tool subcommands and APIs, ensuring that tools uniformly retrieve > > > > event attributes directly from the sample itself. This simplifies > > > > function signatures and improves API consistency. > > > > > > > > Additionally, this series incorporates subsequent memory safety > > > > improvements, robust bounds checking, and memory leak corrections > > > > prompted by address sanitizers and Sashiko code reviews. > > > > > > > > The original combined RFC patch: > > > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > > > > was split up per reviewer requests. > > > > > > > > v14: > > > > - Incorporate Acked-by tags from Namhyung Kim. > > > > - Re-polish and expand commit descriptions to provide thorough explanations > > > > of the changes and their rationale. > > > > > > Thanks, tentatively applying it to perf-tools-next, for v7.2. > > > > Had to apply a minor fixup for a conflict with anoter patch series from > > you, the LBR one. > > Sorry for that and thanks for dealing with the conflict! Any chance > you could push to tmp.perf-tools-next? I have some other series that > will conflict and it would be nice to rebase those changes there. I just removed the last two patches, the ones for kwork, as: + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= NO_LIBELF=1 -C tools/perf O=/tmp/build/perf make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' BUILD: Doing 'make -j32' parallel build Makefile.config:647: Warning: Disabled BPF skeletons as libelf is required by bpftool Makefile.config:745: Disabling post unwind, no support found. Makefile.config:856: Missing python setuptools, the python binding won't be built, please install python3-setuptools or equivalent Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. Auto-detecting system features: ... libdw: [ on ] ... glibc: [ on ] ... libelf: [ on ] ... libnuma: [ on ] ... numa_num_possible_cpus: [ on ] ... libpython: [ on ] ... libcapstone: [ on ] ... llvm-perf: [ on ] ... zlib: [ on ] ... lzma: [ on ] ... bpf: [ on ] ... libaio: [ on ] ... libzstd: [ on ] ... libopenssl: [ on ] ... rust: [ OFF ] CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o <SNIP> LINK /tmp/build/perf/perf /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_pop_atom': builtin-kwork.c:(.text+0x59c92): undefined reference to `work_exit' /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_push_atom.constprop.0': builtin-kwork.c:(.text+0x5aba8): undefined reference to `work_exit' /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `cmd_kwork': (.text+0x5bb4b): undefined reference to `work_exit' /usr/bin/ld: (.text+0x5bbae): undefined reference to `work_exit' /usr/bin/ld: (.text+0x5c3b2): undefined reference to `work_exit' collect2: error: ld returned 1 exit status make[2]: *** [Makefile.perf:586: /tmp/build/perf/perf] Error 1 make[1]: *** [Makefile.perf:289: sub-make] Error 2 make: *** [Makefile:76: all] Error 2 make: Leaving directory '/git/perf-7.1.0-rc2/tools/perf' + exit 1 toolsbuilder@five:~$ I'm doing some more tests, just a moment and I'll push what I have to tmp.perf-tools-next. - Arnaldo
On Wed, May 20, 2026 at 05:49:57PM -0300, Arnaldo Carvalho de Melo wrote:
> On Wed, May 20, 2026 at 01:40:48PM -0700, Ian Rogers wrote:
> > On Wed, May 20, 2026 at 12:46 PM Arnaldo Carvalho de Melo
> > <acme@kernel.org> wrote:
> > >
> > > On Wed, May 20, 2026 at 04:15:30PM -0300, Arnaldo Carvalho de Melo wrote:
> > > > On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote:
> > > > > Nearly all perf code ends up passing an evsel with the perf_sample,
> > > > > which is problematic if you want to rewrite the evsel (such as with
> > > > > off-CPU processing) because all uses of the evsel need to be fixed
> > > > > up. Since the perf_sample now carries its own resolved evsel pointer,
> > > > > passing evsel separately is redundant and error-prone (allowing
> > > > > parameter divergence bugs).
> > > > >
> > > > > This series cleans up the redundant evsel parameter across all perf
> > > > > tool subcommands and APIs, ensuring that tools uniformly retrieve
> > > > > event attributes directly from the sample itself. This simplifies
> > > > > function signatures and improves API consistency.
> > > > >
> > > > > Additionally, this series incorporates subsequent memory safety
> > > > > improvements, robust bounds checking, and memory leak corrections
> > > > > prompted by address sanitizers and Sashiko code reviews.
> > > > >
> > > > > The original combined RFC patch:
> > > > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/
> > > > > was split up per reviewer requests.
> > > > >
> > > > > v14:
> > > > > - Incorporate Acked-by tags from Namhyung Kim.
> > > > > - Re-polish and expand commit descriptions to provide thorough explanations
> > > > > of the changes and their rationale.
> > > >
> > > > Thanks, tentatively applying it to perf-tools-next, for v7.2.
> > >
> > > Had to apply a minor fixup for a conflict with anoter patch series from
> > > you, the LBR one.
> >
> > Sorry for that and thanks for dealing with the conflict! Any chance
> > you could push to tmp.perf-tools-next? I have some other series that
> > will conflict and it would be nice to rebase those changes there.
>
> I just removed the last two patches, the ones for kwork, as:
>
> + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= NO_LIBELF=1 -C tools/perf O=/tmp/build/perf
> make: Entering directory '/git/perf-7.1.0-rc2/tools/perf'
> BUILD: Doing 'make -j32' parallel build
> Makefile.config:647: Warning: Disabled BPF skeletons as libelf is required by bpftool
> Makefile.config:745: Disabling post unwind, no support found.
> Makefile.config:856: Missing python setuptools, the python binding won't be built, please install python3-setuptools or equivalent
> Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev
> Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev
> Makefile.config:1135: Rust is not found. Test workloads with rust are disabled.
>
> Auto-detecting system features:
> ... libdw: [ on ]
> ... glibc: [ on ]
> ... libelf: [ on ]
> ... libnuma: [ on ]
> ... numa_num_possible_cpus: [ on ]
> ... libpython: [ on ]
> ... libcapstone: [ on ]
> ... llvm-perf: [ on ]
> ... zlib: [ on ]
> ... lzma: [ on ]
> ... bpf: [ on ]
> ... libaio: [ on ]
> ... libzstd: [ on ]
> ... libopenssl: [ on ]
> ... rust: [ OFF ]
>
> CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o
> <SNIP>
> LINK /tmp/build/perf/perf
> /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_pop_atom':
> builtin-kwork.c:(.text+0x59c92): undefined reference to `work_exit'
> /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_push_atom.constprop.0':
> builtin-kwork.c:(.text+0x5aba8): undefined reference to `work_exit'
> /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `cmd_kwork':
> (.text+0x5bb4b): undefined reference to `work_exit'
> /usr/bin/ld: (.text+0x5bbae): undefined reference to `work_exit'
> /usr/bin/ld: (.text+0x5c3b2): undefined reference to `work_exit'
> collect2: error: ld returned 1 exit status
> make[2]: *** [Makefile.perf:586: /tmp/build/perf/perf] Error 1
> make[1]: *** [Makefile.perf:289: sub-make] Error 2
> make: *** [Makefile:76: all] Error 2
> make: Leaving directory '/git/perf-7.1.0-rc2/tools/perf'
> + exit 1
> toolsbuilder@five:~$
>
>
> I'm doing some more tests, just a moment and I'll push what I have to
> tmp.perf-tools-next.
Also:
toolsbuilder@five:~$ cat dm.log/almalinux\:8
<SNIP>
OLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/8/lto-wrapper
OFFLOAD_TARGET_NAMES=nvptx-none
OFFLOAD_TARGET_DEFAULT=1
Target: x86_64-redhat-linux
Configured with: ../configure --enable-bootstrap --enable-languages=c,c++,fortran,lto --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugs.almalinux.org/ --enable-shared --enable-threads=posix --enable-checking=release --enable-multilib --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-gcc-major-version-only --with-linker-hash-style=gnu --enable-plugin --enable-initfini-array --with-isl --disable-libmpx --enable-offload-targets=nvptx-none --without-cuda-driver --enable-gnu-indirect-function --enable-cet --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux
Thread model: posix
gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC)
+ make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= -C tools/perf O=/tmp/build/perf
make: Entering directory '/git/perf-7.1.0-rc2/tools/perf'
BUILD: Doing 'make -j32' parallel build
Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev
Makefile.config:1061: No libcapstone found, disables disasm engine support for 'perf script', please install libcapstone-dev/capstone-devel
Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev
Makefile.config:1135: Rust is not found. Test workloads with rust are disabled.
Auto-detecting system features:
... libdw: [ on ]
... glibc: [ on ]
... libelf: [ on ]
... libnuma: [ on ]
... numa_num_possible_cpus: [ on ]
... libpython: [ on ]
... libcapstone: [ OFF ]
... llvm-perf: [ on ]
... zlib: [ on ]
... lzma: [ on ]
... bpf: [ on ]
... libaio: [ on ]
... libzstd: [ on ]
... libopenssl: [ on ]
... rust: [ OFF ]
CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o
CC /tmp/build/perf/dlfilters/dlfilter-test-api-v2.o
<SNIP>
COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/cache.json
COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/counter.json
COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/floating-point.json
COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/frontend.json
COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/memory.json
COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/other.json
COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/pipeline.json
COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/virtual-memory.json
GEN /tmp/build/perf/pmu-events/test-empty-pmu-events.c
Traceback (most recent call last):
File "pmu-events/jevents.py", line 1524, in <module>
main()
File "pmu-events/jevents.py", line 1487, in main
with concurrent.futures.ProcessPoolExecutor(initializer=_init_worker, initargs=(_arch_std_events,)) as executor:
TypeError: __init__() got an unexpected keyword argument 'initializer'
make[3]: *** [pmu-events/Build:176: /tmp/build/perf/pmu-events/test-empty-pmu-events.c] Error 1
make[3]: *** Deleting file '/tmp/build/perf/pmu-events/test-empty-pmu-events.c'
make[2]: *** [Makefile.perf:554: /tmp/build/perf/pmu-events/pmu-events-in.o] Error 2
make[2]: *** Waiting for unfinished jobs....
On Wed, May 20, 2026 at 1:51 PM Arnaldo Carvalho de Melo <acme@kernel.org> wrote: > > On Wed, May 20, 2026 at 05:49:57PM -0300, Arnaldo Carvalho de Melo wrote: > > On Wed, May 20, 2026 at 01:40:48PM -0700, Ian Rogers wrote: > > > On Wed, May 20, 2026 at 12:46 PM Arnaldo Carvalho de Melo > > > <acme@kernel.org> wrote: > > > > > > > > On Wed, May 20, 2026 at 04:15:30PM -0300, Arnaldo Carvalho de Melo wrote: > > > > > On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote: > > > > > > Nearly all perf code ends up passing an evsel with the perf_sample, > > > > > > which is problematic if you want to rewrite the evsel (such as with > > > > > > off-CPU processing) because all uses of the evsel need to be fixed > > > > > > up. Since the perf_sample now carries its own resolved evsel pointer, > > > > > > passing evsel separately is redundant and error-prone (allowing > > > > > > parameter divergence bugs). > > > > > > > > > > > > This series cleans up the redundant evsel parameter across all perf > > > > > > tool subcommands and APIs, ensuring that tools uniformly retrieve > > > > > > event attributes directly from the sample itself. This simplifies > > > > > > function signatures and improves API consistency. > > > > > > > > > > > > Additionally, this series incorporates subsequent memory safety > > > > > > improvements, robust bounds checking, and memory leak corrections > > > > > > prompted by address sanitizers and Sashiko code reviews. > > > > > > > > > > > > The original combined RFC patch: > > > > > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > > > > > > was split up per reviewer requests. > > > > > > > > > > > > v14: > > > > > > - Incorporate Acked-by tags from Namhyung Kim. > > > > > > - Re-polish and expand commit descriptions to provide thorough explanations > > > > > > of the changes and their rationale. > > > > > > > > > > Thanks, tentatively applying it to perf-tools-next, for v7.2. > > > > > > > > Had to apply a minor fixup for a conflict with anoter patch series from > > > > you, the LBR one. > > > > > > Sorry for that and thanks for dealing with the conflict! Any chance > > > you could push to tmp.perf-tools-next? I have some other series that > > > will conflict and it would be nice to rebase those changes there. > > > > I just removed the last two patches, the ones for kwork, as: > > > > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= NO_LIBELF=1 -C tools/perf O=/tmp/build/perf > > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' > > BUILD: Doing 'make -j32' parallel build > > Makefile.config:647: Warning: Disabled BPF skeletons as libelf is required by bpftool > > Makefile.config:745: Disabling post unwind, no support found. > > Makefile.config:856: Missing python setuptools, the python binding won't be built, please install python3-setuptools or equivalent > > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev > > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev > > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. > > > > Auto-detecting system features: > > ... libdw: [ on ] > > ... glibc: [ on ] > > ... libelf: [ on ] > > ... libnuma: [ on ] > > ... numa_num_possible_cpus: [ on ] > > ... libpython: [ on ] > > ... libcapstone: [ on ] > > ... llvm-perf: [ on ] > > ... zlib: [ on ] > > ... lzma: [ on ] > > ... bpf: [ on ] > > ... libaio: [ on ] > > ... libzstd: [ on ] > > ... libopenssl: [ on ] > > ... rust: [ OFF ] > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o > > <SNIP> > > LINK /tmp/build/perf/perf > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_pop_atom': > > builtin-kwork.c:(.text+0x59c92): undefined reference to `work_exit' > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_push_atom.constprop.0': > > builtin-kwork.c:(.text+0x5aba8): undefined reference to `work_exit' > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `cmd_kwork': > > (.text+0x5bb4b): undefined reference to `work_exit' > > /usr/bin/ld: (.text+0x5bbae): undefined reference to `work_exit' > > /usr/bin/ld: (.text+0x5c3b2): undefined reference to `work_exit' > > collect2: error: ld returned 1 exit status > > make[2]: *** [Makefile.perf:586: /tmp/build/perf/perf] Error 1 > > make[1]: *** [Makefile.perf:289: sub-make] Error 2 > > make: *** [Makefile:76: all] Error 2 > > make: Leaving directory '/git/perf-7.1.0-rc2/tools/perf' > > + exit 1 > > toolsbuilder@five:~$ > > > > > > I'm doing some more tests, just a moment and I'll push what I have to > > tmp.perf-tools-next. > > Also: > > toolsbuilder@five:~$ cat dm.log/almalinux\:8 > <SNIP> > OLLECT_GCC=gcc > COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/8/lto-wrapper > OFFLOAD_TARGET_NAMES=nvptx-none > OFFLOAD_TARGET_DEFAULT=1 > Target: x86_64-redhat-linux > Configured with: ../configure --enable-bootstrap --enable-languages=c,c++,fortran,lto --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugs.almalinux.org/ --enable-shared --enable-threads=posix --enable-checking=release --enable-multilib --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-gcc-major-version-only --with-linker-hash-style=gnu --enable-plugin --enable-initfini-array --with-isl --disable-libmpx --enable-offload-targets=nvptx-none --without-cuda-driver --enable-gnu-indirect-function --enable-cet --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux > Thread model: posix > gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= -C tools/perf O=/tmp/build/perf > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' > BUILD: Doing 'make -j32' parallel build > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev > Makefile.config:1061: No libcapstone found, disables disasm engine support for 'perf script', please install libcapstone-dev/capstone-devel > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. > > Auto-detecting system features: > ... libdw: [ on ] > ... glibc: [ on ] > ... libelf: [ on ] > ... libnuma: [ on ] > ... numa_num_possible_cpus: [ on ] > ... libpython: [ on ] > ... libcapstone: [ OFF ] > ... llvm-perf: [ on ] > ... zlib: [ on ] > ... lzma: [ on ] > ... bpf: [ on ] > ... libaio: [ on ] > ... libzstd: [ on ] > ... libopenssl: [ on ] > ... rust: [ OFF ] > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v2.o > <SNIP> > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/cache.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/counter.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/floating-point.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/frontend.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/memory.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/other.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/pipeline.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/virtual-memory.json > GEN /tmp/build/perf/pmu-events/test-empty-pmu-events.c > Traceback (most recent call last): > File "pmu-events/jevents.py", line 1524, in <module> > main() > File "pmu-events/jevents.py", line 1487, in main > with concurrent.futures.ProcessPoolExecutor(initializer=_init_worker, initargs=(_arch_std_events,)) as executor: > TypeError: __init__() got an unexpected keyword argument 'initializer' The `initializer` and `initargs` parameters were added to `concurrent.futures.ProcessPoolExecutor`.futures.ProcessPoolExecutor in Python 3.7. Python 3.7 is past its end of life and the kernel build specified that Python 3.9 is the minimum supported: https://docs.kernel.org/process/changes.html A fix for this is to disable jevents for older Python versions, but detecting this will incur a build-time cost. Thanks, Ian > make[3]: *** [pmu-events/Build:176: /tmp/build/perf/pmu-events/test-empty-pmu-events.c] Error 1 > make[3]: *** Deleting file '/tmp/build/perf/pmu-events/test-empty-pmu-events.c' > make[2]: *** [Makefile.perf:554: /tmp/build/perf/pmu-events/pmu-events-in.o] Error 2 > make[2]: *** Waiting for unfinished jobs.... >
On Wed, May 20, 2026 at 05:51:36PM -0300, Arnaldo Carvalho de Melo wrote: > On Wed, May 20, 2026 at 05:49:57PM -0300, Arnaldo Carvalho de Melo wrote: > > On Wed, May 20, 2026 at 01:40:48PM -0700, Ian Rogers wrote: > > > On Wed, May 20, 2026 at 12:46 PM Arnaldo Carvalho de Melo > > > <acme@kernel.org> wrote: > > > > > > > > On Wed, May 20, 2026 at 04:15:30PM -0300, Arnaldo Carvalho de Melo wrote: > > > > > On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote: > > > > > > Nearly all perf code ends up passing an evsel with the perf_sample, > > > > > > which is problematic if you want to rewrite the evsel (such as with > > > > > > off-CPU processing) because all uses of the evsel need to be fixed > > > > > > up. Since the perf_sample now carries its own resolved evsel pointer, > > > > > > passing evsel separately is redundant and error-prone (allowing > > > > > > parameter divergence bugs). > > > > > > > > > > > > This series cleans up the redundant evsel parameter across all perf > > > > > > tool subcommands and APIs, ensuring that tools uniformly retrieve > > > > > > event attributes directly from the sample itself. This simplifies > > > > > > function signatures and improves API consistency. > > > > > > > > > > > > Additionally, this series incorporates subsequent memory safety > > > > > > improvements, robust bounds checking, and memory leak corrections > > > > > > prompted by address sanitizers and Sashiko code reviews. > > > > > > > > > > > > The original combined RFC patch: > > > > > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > > > > > > was split up per reviewer requests. > > > > > > > > > > > > v14: > > > > > > - Incorporate Acked-by tags from Namhyung Kim. > > > > > > - Re-polish and expand commit descriptions to provide thorough explanations > > > > > > of the changes and their rationale. > > > > > > > > > > Thanks, tentatively applying it to perf-tools-next, for v7.2. > > > > > > > > Had to apply a minor fixup for a conflict with anoter patch series from > > > > you, the LBR one. > > > > > > Sorry for that and thanks for dealing with the conflict! Any chance > > > you could push to tmp.perf-tools-next? I have some other series that > > > will conflict and it would be nice to rebase those changes there. > > > > I just removed the last two patches, the ones for kwork, as: > > > > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= NO_LIBELF=1 -C tools/perf O=/tmp/build/perf > > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' > > BUILD: Doing 'make -j32' parallel build > > Makefile.config:647: Warning: Disabled BPF skeletons as libelf is required by bpftool > > Makefile.config:745: Disabling post unwind, no support found. > > Makefile.config:856: Missing python setuptools, the python binding won't be built, please install python3-setuptools or equivalent > > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev > > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev > > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. > > > > Auto-detecting system features: > > ... libdw: [ on ] > > ... glibc: [ on ] > > ... libelf: [ on ] > > ... libnuma: [ on ] > > ... numa_num_possible_cpus: [ on ] > > ... libpython: [ on ] > > ... libcapstone: [ on ] > > ... llvm-perf: [ on ] > > ... zlib: [ on ] > > ... lzma: [ on ] > > ... bpf: [ on ] > > ... libaio: [ on ] > > ... libzstd: [ on ] > > ... libopenssl: [ on ] > > ... rust: [ OFF ] > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o > > <SNIP> > > LINK /tmp/build/perf/perf > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_pop_atom': > > builtin-kwork.c:(.text+0x59c92): undefined reference to `work_exit' > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_push_atom.constprop.0': > > builtin-kwork.c:(.text+0x5aba8): undefined reference to `work_exit' > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `cmd_kwork': > > (.text+0x5bb4b): undefined reference to `work_exit' > > /usr/bin/ld: (.text+0x5bbae): undefined reference to `work_exit' > > /usr/bin/ld: (.text+0x5c3b2): undefined reference to `work_exit' > > collect2: error: ld returned 1 exit status > > make[2]: *** [Makefile.perf:586: /tmp/build/perf/perf] Error 1 > > make[1]: *** [Makefile.perf:289: sub-make] Error 2 > > make: *** [Makefile:76: all] Error 2 > > make: Leaving directory '/git/perf-7.1.0-rc2/tools/perf' > > + exit 1 > > toolsbuilder@five:~$ > > > > > > I'm doing some more tests, just a moment and I'll push what I have to > > tmp.perf-tools-next. > > Also: > > toolsbuilder@five:~$ cat dm.log/almalinux\:8 > <SNIP> > OLLECT_GCC=gcc > COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/8/lto-wrapper > OFFLOAD_TARGET_NAMES=nvptx-none > OFFLOAD_TARGET_DEFAULT=1 > Target: x86_64-redhat-linux > Configured with: ../configure --enable-bootstrap --enable-languages=c,c++,fortran,lto --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugs.almalinux.org/ --enable-shared --enable-threads=posix --enable-checking=release --enable-multilib --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-gcc-major-version-only --with-linker-hash-style=gnu --enable-plugin --enable-initfini-array --with-isl --disable-libmpx --enable-offload-targets=nvptx-none --without-cuda-driver --enable-gnu-indirect-function --enable-cet --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux > Thread model: posix > gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= -C tools/perf O=/tmp/build/perf > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' > BUILD: Doing 'make -j32' parallel build > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev > Makefile.config:1061: No libcapstone found, disables disasm engine support for 'perf script', please install libcapstone-dev/capstone-devel > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. > > Auto-detecting system features: > ... libdw: [ on ] > ... glibc: [ on ] > ... libelf: [ on ] > ... libnuma: [ on ] > ... numa_num_possible_cpus: [ on ] > ... libpython: [ on ] > ... libcapstone: [ OFF ] > ... llvm-perf: [ on ] > ... zlib: [ on ] > ... lzma: [ on ] > ... bpf: [ on ] > ... libaio: [ on ] > ... libzstd: [ on ] > ... libopenssl: [ on ] > ... rust: [ OFF ] > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v2.o > <SNIP> > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/cache.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/counter.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/floating-point.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/frontend.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/memory.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/other.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/pipeline.json > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/virtual-memory.json > GEN /tmp/build/perf/pmu-events/test-empty-pmu-events.c > Traceback (most recent call last): > File "pmu-events/jevents.py", line 1524, in <module> > main() > File "pmu-events/jevents.py", line 1487, in main > with concurrent.futures.ProcessPoolExecutor(initializer=_init_worker, initargs=(_arch_std_events,)) as executor: > TypeError: __init__() got an unexpected keyword argument 'initializer' > make[3]: *** [pmu-events/Build:176: /tmp/build/perf/pmu-events/test-empty-pmu-events.c] Error 1 > make[3]: *** Deleting file '/tmp/build/perf/pmu-events/test-empty-pmu-events.c' > make[2]: *** [Makefile.perf:554: /tmp/build/perf/pmu-events/pmu-events-in.o] Error 2 > make[2]: *** Waiting for unfinished jobs.... But: toolsbuilder@five:~$ export BUILD_TARBALL=http://192.168.86.5/perf/perf-7.1.0-rc2.tar.xz toolsbuilder@five:~$ time dm 1 8.77 almalinux:8 : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) 2 84.96 almalinux:9 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-11) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-3.el9) flex 2.6.4 3: 103.23 almalinux:9-i386 : Ok gcc (GCC) 11.4.1 20231218 (Red Hat 11.4.1-3) , clang version 17.0.6 (AlmaLinux OS Foundation 17.0.6-5.el9) flex 2.6.4 4 90.91 almalinux:10 : Ok gcc (GCC) 14.3.1 20250617 (Red Hat 14.3.1-2) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-1.el10.alma.1) flex 2.6.4 5 118.21 alpine:3.16 : Ok gcc (Alpine 11.2.1_git20220219) 11.2.1 20220219 , Alpine clang version 13.0.1 flex 2.6.4 ...
On Wed, May 20, 2026 at 1:55 PM Arnaldo Carvalho de Melo <acme@kernel.org> wrote: > > On Wed, May 20, 2026 at 05:51:36PM -0300, Arnaldo Carvalho de Melo wrote: > > On Wed, May 20, 2026 at 05:49:57PM -0300, Arnaldo Carvalho de Melo wrote: > > > On Wed, May 20, 2026 at 01:40:48PM -0700, Ian Rogers wrote: > > > > On Wed, May 20, 2026 at 12:46 PM Arnaldo Carvalho de Melo > > > > <acme@kernel.org> wrote: > > > > > > > > > > On Wed, May 20, 2026 at 04:15:30PM -0300, Arnaldo Carvalho de Melo wrote: > > > > > > On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote: > > > > > > > Nearly all perf code ends up passing an evsel with the perf_sample, > > > > > > > which is problematic if you want to rewrite the evsel (such as with > > > > > > > off-CPU processing) because all uses of the evsel need to be fixed > > > > > > > up. Since the perf_sample now carries its own resolved evsel pointer, > > > > > > > passing evsel separately is redundant and error-prone (allowing > > > > > > > parameter divergence bugs). > > > > > > > > > > > > > > This series cleans up the redundant evsel parameter across all perf > > > > > > > tool subcommands and APIs, ensuring that tools uniformly retrieve > > > > > > > event attributes directly from the sample itself. This simplifies > > > > > > > function signatures and improves API consistency. > > > > > > > > > > > > > > Additionally, this series incorporates subsequent memory safety > > > > > > > improvements, robust bounds checking, and memory leak corrections > > > > > > > prompted by address sanitizers and Sashiko code reviews. > > > > > > > > > > > > > > The original combined RFC patch: > > > > > > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > > > > > > > was split up per reviewer requests. > > > > > > > > > > > > > > v14: > > > > > > > - Incorporate Acked-by tags from Namhyung Kim. > > > > > > > - Re-polish and expand commit descriptions to provide thorough explanations > > > > > > > of the changes and their rationale. > > > > > > > > > > > > Thanks, tentatively applying it to perf-tools-next, for v7.2. > > > > > > > > > > Had to apply a minor fixup for a conflict with anoter patch series from > > > > > you, the LBR one. > > > > > > > > Sorry for that and thanks for dealing with the conflict! Any chance > > > > you could push to tmp.perf-tools-next? I have some other series that > > > > will conflict and it would be nice to rebase those changes there. > > > > > > I just removed the last two patches, the ones for kwork, as: > > > > > > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= NO_LIBELF=1 -C tools/perf O=/tmp/build/perf > > > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' > > > BUILD: Doing 'make -j32' parallel build > > > Makefile.config:647: Warning: Disabled BPF skeletons as libelf is required by bpftool > > > Makefile.config:745: Disabling post unwind, no support found. > > > Makefile.config:856: Missing python setuptools, the python binding won't be built, please install python3-setuptools or equivalent > > > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev > > > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev > > > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. > > > > > > Auto-detecting system features: > > > ... libdw: [ on ] > > > ... glibc: [ on ] > > > ... libelf: [ on ] > > > ... libnuma: [ on ] > > > ... numa_num_possible_cpus: [ on ] > > > ... libpython: [ on ] > > > ... libcapstone: [ on ] > > > ... llvm-perf: [ on ] > > > ... zlib: [ on ] > > > ... lzma: [ on ] > > > ... bpf: [ on ] > > > ... libaio: [ on ] > > > ... libzstd: [ on ] > > > ... libopenssl: [ on ] > > > ... rust: [ OFF ] > > > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o > > > <SNIP> > > > LINK /tmp/build/perf/perf > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_pop_atom': > > > builtin-kwork.c:(.text+0x59c92): undefined reference to `work_exit' > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_push_atom.constprop.0': > > > builtin-kwork.c:(.text+0x5aba8): undefined reference to `work_exit' > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `cmd_kwork': > > > (.text+0x5bb4b): undefined reference to `work_exit' > > > /usr/bin/ld: (.text+0x5bbae): undefined reference to `work_exit' > > > /usr/bin/ld: (.text+0x5c3b2): undefined reference to `work_exit' > > > collect2: error: ld returned 1 exit status > > > make[2]: *** [Makefile.perf:586: /tmp/build/perf/perf] Error 1 > > > make[1]: *** [Makefile.perf:289: sub-make] Error 2 > > > make: *** [Makefile:76: all] Error 2 > > > make: Leaving directory '/git/perf-7.1.0-rc2/tools/perf' > > > + exit 1 > > > toolsbuilder@five:~$ > > > > > > > > > I'm doing some more tests, just a moment and I'll push what I have to > > > tmp.perf-tools-next. > > > > Also: > > > > toolsbuilder@five:~$ cat dm.log/almalinux\:8 > > <SNIP> > > OLLECT_GCC=gcc > > COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/8/lto-wrapper > > OFFLOAD_TARGET_NAMES=nvptx-none > > OFFLOAD_TARGET_DEFAULT=1 > > Target: x86_64-redhat-linux > > Configured with: ../configure --enable-bootstrap --enable-languages=c,c++,fortran,lto --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugs.almalinux.org/ --enable-shared --enable-threads=posix --enable-checking=release --enable-multilib --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-gcc-major-version-only --with-linker-hash-style=gnu --enable-plugin --enable-initfini-array --with-isl --disable-libmpx --enable-offload-targets=nvptx-none --without-cuda-driver --enable-gnu-indirect-function --enable-cet --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux > > Thread model: posix > > gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= -C tools/perf O=/tmp/build/perf > > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' > > BUILD: Doing 'make -j32' parallel build > > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev > > Makefile.config:1061: No libcapstone found, disables disasm engine support for 'perf script', please install libcapstone-dev/capstone-devel > > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev > > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. > > > > Auto-detecting system features: > > ... libdw: [ on ] > > ... glibc: [ on ] > > ... libelf: [ on ] > > ... libnuma: [ on ] > > ... numa_num_possible_cpus: [ on ] > > ... libpython: [ on ] > > ... libcapstone: [ OFF ] > > ... llvm-perf: [ on ] > > ... zlib: [ on ] > > ... lzma: [ on ] > > ... bpf: [ on ] > > ... libaio: [ on ] > > ... libzstd: [ on ] > > ... libopenssl: [ on ] > > ... rust: [ OFF ] > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v2.o > > <SNIP> > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/cache.json > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/counter.json > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/floating-point.json > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/frontend.json > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/memory.json > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/other.json > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/pipeline.json > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/virtual-memory.json > > GEN /tmp/build/perf/pmu-events/test-empty-pmu-events.c > > Traceback (most recent call last): > > File "pmu-events/jevents.py", line 1524, in <module> > > main() > > File "pmu-events/jevents.py", line 1487, in main > > with concurrent.futures.ProcessPoolExecutor(initializer=_init_worker, initargs=(_arch_std_events,)) as executor: > > TypeError: __init__() got an unexpected keyword argument 'initializer' > > make[3]: *** [pmu-events/Build:176: /tmp/build/perf/pmu-events/test-empty-pmu-events.c] Error 1 > > make[3]: *** Deleting file '/tmp/build/perf/pmu-events/test-empty-pmu-events.c' > > make[2]: *** [Makefile.perf:554: /tmp/build/perf/pmu-events/pmu-events-in.o] Error 2 > > make[2]: *** Waiting for unfinished jobs.... > > But: > > toolsbuilder@five:~$ export BUILD_TARBALL=http://192.168.86.5/perf/perf-7.1.0-rc2.tar.xz > toolsbuilder@five:~$ time dm > 1 8.77 almalinux:8 : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > 2 84.96 almalinux:9 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-11) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-3.el9) flex 2.6.4 > 3: 103.23 almalinux:9-i386 : Ok gcc (GCC) 11.4.1 20231218 (Red Hat 11.4.1-3) , clang version 17.0.6 (AlmaLinux OS Foundation 17.0.6-5.el9) flex 2.6.4 > 4 90.91 almalinux:10 : Ok gcc (GCC) 14.3.1 20250617 (Red Hat 14.3.1-2) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-1.el10.alma.1) flex 2.6.4 > 5 118.21 alpine:3.16 : Ok gcc (Alpine 11.2.1_git20220219) 11.2.1 20220219 , Alpine clang version 13.0.1 flex 2.6.4 > ... You can drop the parallelize jevent.py changes and I'll follow up. Sorry for the breakage. Thanks, Ian
On Wed, May 20, 2026 at 02:01:03PM -0700, Ian Rogers wrote:
> On Wed, May 20, 2026 at 1:55 PM Arnaldo Carvalho de Melo
> <acme@kernel.org> wrote:
> >
> > On Wed, May 20, 2026 at 05:51:36PM -0300, Arnaldo Carvalho de Melo wrote:
> > > On Wed, May 20, 2026 at 05:49:57PM -0300, Arnaldo Carvalho de Melo wrote:
> > > > On Wed, May 20, 2026 at 01:40:48PM -0700, Ian Rogers wrote:
> > > > > On Wed, May 20, 2026 at 12:46 PM Arnaldo Carvalho de Melo
> > > > > <acme@kernel.org> wrote:
> > > > > >
> > > > > > On Wed, May 20, 2026 at 04:15:30PM -0300, Arnaldo Carvalho de Melo wrote:
> > > > > > > On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote:
> > > > > > > > Nearly all perf code ends up passing an evsel with the perf_sample,
> > > > > > > > which is problematic if you want to rewrite the evsel (such as with
> > > > > > > > off-CPU processing) because all uses of the evsel need to be fixed
> > > > > > > > up. Since the perf_sample now carries its own resolved evsel pointer,
> > > > > > > > passing evsel separately is redundant and error-prone (allowing
> > > > > > > > parameter divergence bugs).
> > > > > > > >
> > > > > > > > This series cleans up the redundant evsel parameter across all perf
> > > > > > > > tool subcommands and APIs, ensuring that tools uniformly retrieve
> > > > > > > > event attributes directly from the sample itself. This simplifies
> > > > > > > > function signatures and improves API consistency.
> > > > > > > >
> > > > > > > > Additionally, this series incorporates subsequent memory safety
> > > > > > > > improvements, robust bounds checking, and memory leak corrections
> > > > > > > > prompted by address sanitizers and Sashiko code reviews.
> > > > > > > >
> > > > > > > > The original combined RFC patch:
> > > > > > > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/
> > > > > > > > was split up per reviewer requests.
> > > > > > > >
> > > > > > > > v14:
> > > > > > > > - Incorporate Acked-by tags from Namhyung Kim.
> > > > > > > > - Re-polish and expand commit descriptions to provide thorough explanations
> > > > > > > > of the changes and their rationale.
> > > > > > >
> > > > > > > Thanks, tentatively applying it to perf-tools-next, for v7.2.
> > > > > >
> > > > > > Had to apply a minor fixup for a conflict with anoter patch series from
> > > > > > you, the LBR one.
> > > > >
> > > > > Sorry for that and thanks for dealing with the conflict! Any chance
> > > > > you could push to tmp.perf-tools-next? I have some other series that
> > > > > will conflict and it would be nice to rebase those changes there.
> > > >
> > > > I just removed the last two patches, the ones for kwork, as:
> > > >
> > > > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= NO_LIBELF=1 -C tools/perf O=/tmp/build/perf
> > > > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf'
> > > > BUILD: Doing 'make -j32' parallel build
> > > > Makefile.config:647: Warning: Disabled BPF skeletons as libelf is required by bpftool
> > > > Makefile.config:745: Disabling post unwind, no support found.
> > > > Makefile.config:856: Missing python setuptools, the python binding won't be built, please install python3-setuptools or equivalent
> > > > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev
> > > > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev
> > > > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled.
> > > >
> > > > Auto-detecting system features:
> > > > ... libdw: [ on ]
> > > > ... glibc: [ on ]
> > > > ... libelf: [ on ]
> > > > ... libnuma: [ on ]
> > > > ... numa_num_possible_cpus: [ on ]
> > > > ... libpython: [ on ]
> > > > ... libcapstone: [ on ]
> > > > ... llvm-perf: [ on ]
> > > > ... zlib: [ on ]
> > > > ... lzma: [ on ]
> > > > ... bpf: [ on ]
> > > > ... libaio: [ on ]
> > > > ... libzstd: [ on ]
> > > > ... libopenssl: [ on ]
> > > > ... rust: [ OFF ]
> > > >
> > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o
> > > > <SNIP>
> > > > LINK /tmp/build/perf/perf
> > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_pop_atom':
> > > > builtin-kwork.c:(.text+0x59c92): undefined reference to `work_exit'
> > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_push_atom.constprop.0':
> > > > builtin-kwork.c:(.text+0x5aba8): undefined reference to `work_exit'
> > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `cmd_kwork':
> > > > (.text+0x5bb4b): undefined reference to `work_exit'
> > > > /usr/bin/ld: (.text+0x5bbae): undefined reference to `work_exit'
> > > > /usr/bin/ld: (.text+0x5c3b2): undefined reference to `work_exit'
> > > > collect2: error: ld returned 1 exit status
> > > > make[2]: *** [Makefile.perf:586: /tmp/build/perf/perf] Error 1
> > > > make[1]: *** [Makefile.perf:289: sub-make] Error 2
> > > > make: *** [Makefile:76: all] Error 2
> > > > make: Leaving directory '/git/perf-7.1.0-rc2/tools/perf'
> > > > + exit 1
> > > > toolsbuilder@five:~$
> > > >
> > > >
> > > > I'm doing some more tests, just a moment and I'll push what I have to
> > > > tmp.perf-tools-next.
> > >
> > > Also:
> > >
> > > toolsbuilder@five:~$ cat dm.log/almalinux\:8
> > > <SNIP>
> > > OLLECT_GCC=gcc
> > > COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/8/lto-wrapper
> > > OFFLOAD_TARGET_NAMES=nvptx-none
> > > OFFLOAD_TARGET_DEFAULT=1
> > > Target: x86_64-redhat-linux
> > > Configured with: ../configure --enable-bootstrap --enable-languages=c,c++,fortran,lto --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugs.almalinux.org/ --enable-shared --enable-threads=posix --enable-checking=release --enable-multilib --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-gcc-major-version-only --with-linker-hash-style=gnu --enable-plugin --enable-initfini-array --with-isl --disable-libmpx --enable-offload-targets=nvptx-none --without-cuda-driver --enable-gnu-indirect-function --enable-cet --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux
> > > Thread model: posix
> > > gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC)
> > > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= -C tools/perf O=/tmp/build/perf
> > > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf'
> > > BUILD: Doing 'make -j32' parallel build
> > > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev
> > > Makefile.config:1061: No libcapstone found, disables disasm engine support for 'perf script', please install libcapstone-dev/capstone-devel
> > > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev
> > > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled.
> > >
> > > Auto-detecting system features:
> > > ... libdw: [ on ]
> > > ... glibc: [ on ]
> > > ... libelf: [ on ]
> > > ... libnuma: [ on ]
> > > ... numa_num_possible_cpus: [ on ]
> > > ... libpython: [ on ]
> > > ... libcapstone: [ OFF ]
> > > ... llvm-perf: [ on ]
> > > ... zlib: [ on ]
> > > ... lzma: [ on ]
> > > ... bpf: [ on ]
> > > ... libaio: [ on ]
> > > ... libzstd: [ on ]
> > > ... libopenssl: [ on ]
> > > ... rust: [ OFF ]
> > >
> > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o
> > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v2.o
> > > <SNIP>
> > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/cache.json
> > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/counter.json
> > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/floating-point.json
> > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/frontend.json
> > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/memory.json
> > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/other.json
> > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/pipeline.json
> > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/virtual-memory.json
> > > GEN /tmp/build/perf/pmu-events/test-empty-pmu-events.c
> > > Traceback (most recent call last):
> > > File "pmu-events/jevents.py", line 1524, in <module>
> > > main()
> > > File "pmu-events/jevents.py", line 1487, in main
> > > with concurrent.futures.ProcessPoolExecutor(initializer=_init_worker, initargs=(_arch_std_events,)) as executor:
> > > TypeError: __init__() got an unexpected keyword argument 'initializer'
> > > make[3]: *** [pmu-events/Build:176: /tmp/build/perf/pmu-events/test-empty-pmu-events.c] Error 1
> > > make[3]: *** Deleting file '/tmp/build/perf/pmu-events/test-empty-pmu-events.c'
> > > make[2]: *** [Makefile.perf:554: /tmp/build/perf/pmu-events/pmu-events-in.o] Error 2
> > > make[2]: *** Waiting for unfinished jobs....
> >
> > But:
> >
> > toolsbuilder@five:~$ export BUILD_TARBALL=http://192.168.86.5/perf/perf-7.1.0-rc2.tar.xz
> > toolsbuilder@five:~$ time dm
> > 1 8.77 almalinux:8 : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC)
> > 2 84.96 almalinux:9 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-11) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-3.el9) flex 2.6.4
> > 3: 103.23 almalinux:9-i386 : Ok gcc (GCC) 11.4.1 20231218 (Red Hat 11.4.1-3) , clang version 17.0.6 (AlmaLinux OS Foundation 17.0.6-5.el9) flex 2.6.4
> > 4 90.91 almalinux:10 : Ok gcc (GCC) 14.3.1 20250617 (Red Hat 14.3.1-2) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-1.el10.alma.1) flex 2.6.4
> > 5 118.21 alpine:3.16 : Ok gcc (Alpine 11.2.1_git20220219) 11.2.1 20220219 , Alpine clang version 13.0.1 flex 2.6.4
> > ...
>
> You can drop the parallelize jevent.py changes and I'll follow up.
> Sorry for the breakage.
It was just that almalinux:8 so far, its an old system, so I'm keeping
it to see what other systems fail, if any.
So far:
toolsbuilder@five:~$ export BUILD_TARBALL=http://192.168.86.5/perf/perf-7.1.0-rc2.tar.xz
toolsbuilder@five:~$ time dm
1 8.77 almalinux:8 : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC)
2 84.96 almalinux:9 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-11) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-3.el9) flex 2.6.4
3: almalinux:9-i386WARNING: image platform (linux/386) does not match the expected platform (linux/amd64)
WARNING: image platform (linux/386) does not match the expected platform (linux/amd64)
103.23 almalinux:9-i386 : Ok gcc (GCC) 11.4.1 20231218 (Red Hat 11.4.1-3) , clang version 17.0.6 (AlmaLinux OS Foundation 17.0.6-5.el9) flex 2.6.4
4 90.91 almalinux:10 : Ok gcc (GCC) 14.3.1 20250617 (Red Hat 14.3.1-2) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-1.el10.alma.1) flex 2.6.4
5 118.21 alpine:3.16 : Ok gcc (Alpine 11.2.1_git20220219) 11.2.1 20220219 , Alpine clang version 13.0.1 flex 2.6.4
6 106.13 alpine:3.17 : Ok gcc (Alpine 12.2.1_git20220924-r4) 12.2.1 20220924 , Alpine clang version 15.0.7 flex 2.6.4
7 102.80 alpine:3.18 : Ok gcc (Alpine 12.2.1_git20220924-r10) 12.2.1 20220924 , Alpine clang version 16.0.6 flex 2.6.4
8 114.48 alpine:3.19 : Ok gcc (Alpine 13.2.1_git20231014) 13.2.1 20231014 , Alpine clang version 17.0.5 flex 2.6.4
9 109.53 alpine:3.20 : Ok gcc (Alpine 13.2.1_git20240309) 13.2.1 20240309 , Alpine clang version 17.0.6 flex 2.6.4
10 116.20 alpine:3.22 : Ok gcc (Alpine 14.2.0) 14.2.0 , Alpine clang version 20.1.8 flex 2.6.4
11 109.83 alpine:3.23 : Ok gcc (Alpine 15.2.0) 15.2.0 , Alpine clang version 21.1.2 flex 2.6.4
12 119.23 alpine:edge : Ok gcc (Alpine 14.2.0) 14.2.0 , Alpine clang version 19.1.4 flex 2.6.4
13 93.22 amazonlinux:2023 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-5) , clang version 15.0.7 (AWS 15.0.7-3.amzn2023.0.4) flex 2.6.4
14 93.40 amazonlinux:devel : Ok gcc (GCC) 11.3.1 20221121 (Red Hat 11.3.1-4) , clang version 15.0.6 (Amazon Linux 15.0.6-3.amzn2023.0.2) flex 2.6.4
15 102.98 archlinux:base : Ok gcc (GCC) 15.2.1 20250813 , clang version 20.1.8 flex 2.6.4
16 9.58 centos:stream : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-21) (GCC)
17: clearlinux:latest
COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/virtual-memory.json
GEN /tmp/build/perf/pmu-events/test-empty-pmu-events.c
Traceback (most recent call last):
File "pmu-events/jevents.py", line 1524, in <module>
main()
File "pmu-events/jevents.py", line 1487, in main
with concurrent.futures.ProcessPoolExecutor(initializer=_init_worker, initargs=(_arch_std_events,)) as executor:
TypeError: __init__() got an unexpected keyword argument 'initializer'
make[3]: *** [pmu-events/Build:176: /tmp/build/perf/pmu-events/test-empty-pmu-events.c] Error 1
make[3]: *** Deleting file '/tmp/build/perf/pmu-events/test-empty-pmu-events.c'
make[2]: *** [Makefile.perf:554: /tmp/build/perf/pmu-events/pmu-events-in.o] Error 2
make[2]: *** Waiting for unfinished jobs....
Auto-detecting system features:
... clang-bpf-co-re: [ on ]
... llvm: [ on ]
... libcap: [ on ]
... libbfd: [ on ]
MKDIR /tmp/build/perf/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf
centos:stream is the bleeding edge, so I'll drop that as you suggested,
ok.
- Arnaldo
On Wed, May 20, 2026 at 06:13:33PM -0300, Arnaldo Carvalho de Melo wrote: > On Wed, May 20, 2026 at 02:01:03PM -0700, Ian Rogers wrote: > > On Wed, May 20, 2026 at 1:55 PM Arnaldo Carvalho de Melo > > <acme@kernel.org> wrote: > > > > > > On Wed, May 20, 2026 at 05:51:36PM -0300, Arnaldo Carvalho de Melo wrote: > > > > On Wed, May 20, 2026 at 05:49:57PM -0300, Arnaldo Carvalho de Melo wrote: > > > > > On Wed, May 20, 2026 at 01:40:48PM -0700, Ian Rogers wrote: > > > > > > On Wed, May 20, 2026 at 12:46 PM Arnaldo Carvalho de Melo > > > > > > <acme@kernel.org> wrote: > > > > > > > > > > > > > > On Wed, May 20, 2026 at 04:15:30PM -0300, Arnaldo Carvalho de Melo wrote: > > > > > > > > On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote: > > > > > > > > > Nearly all perf code ends up passing an evsel with the perf_sample, > > > > > > > > > which is problematic if you want to rewrite the evsel (such as with > > > > > > > > > off-CPU processing) because all uses of the evsel need to be fixed > > > > > > > > > up. Since the perf_sample now carries its own resolved evsel pointer, > > > > > > > > > passing evsel separately is redundant and error-prone (allowing > > > > > > > > > parameter divergence bugs). > > > > > > > > > > > > > > > > > > This series cleans up the redundant evsel parameter across all perf > > > > > > > > > tool subcommands and APIs, ensuring that tools uniformly retrieve > > > > > > > > > event attributes directly from the sample itself. This simplifies > > > > > > > > > function signatures and improves API consistency. > > > > > > > > > > > > > > > > > > Additionally, this series incorporates subsequent memory safety > > > > > > > > > improvements, robust bounds checking, and memory leak corrections > > > > > > > > > prompted by address sanitizers and Sashiko code reviews. > > > > > > > > > > > > > > > > > > The original combined RFC patch: > > > > > > > > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > > > > > > > > > was split up per reviewer requests. > > > > > > > > > > > > > > > > > > v14: > > > > > > > > > - Incorporate Acked-by tags from Namhyung Kim. > > > > > > > > > - Re-polish and expand commit descriptions to provide thorough explanations > > > > > > > > > of the changes and their rationale. > > > > > > > > > > > > > > > > Thanks, tentatively applying it to perf-tools-next, for v7.2. > > > > > > > > > > > > > > Had to apply a minor fixup for a conflict with anoter patch series from > > > > > > > you, the LBR one. > > > > > > > > > > > > Sorry for that and thanks for dealing with the conflict! Any chance > > > > > > you could push to tmp.perf-tools-next? I have some other series that > > > > > > will conflict and it would be nice to rebase those changes there. > > > > > > > > > > I just removed the last two patches, the ones for kwork, as: > > > > > > > > > > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= NO_LIBELF=1 -C tools/perf O=/tmp/build/perf > > > > > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' > > > > > BUILD: Doing 'make -j32' parallel build > > > > > Makefile.config:647: Warning: Disabled BPF skeletons as libelf is required by bpftool > > > > > Makefile.config:745: Disabling post unwind, no support found. > > > > > Makefile.config:856: Missing python setuptools, the python binding won't be built, please install python3-setuptools or equivalent > > > > > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev > > > > > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev > > > > > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. > > > > > > > > > > Auto-detecting system features: > > > > > ... libdw: [ on ] > > > > > ... glibc: [ on ] > > > > > ... libelf: [ on ] > > > > > ... libnuma: [ on ] > > > > > ... numa_num_possible_cpus: [ on ] > > > > > ... libpython: [ on ] > > > > > ... libcapstone: [ on ] > > > > > ... llvm-perf: [ on ] > > > > > ... zlib: [ on ] > > > > > ... lzma: [ on ] > > > > > ... bpf: [ on ] > > > > > ... libaio: [ on ] > > > > > ... libzstd: [ on ] > > > > > ... libopenssl: [ on ] > > > > > ... rust: [ OFF ] > > > > > > > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o > > > > > <SNIP> > > > > > LINK /tmp/build/perf/perf > > > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_pop_atom': > > > > > builtin-kwork.c:(.text+0x59c92): undefined reference to `work_exit' > > > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_push_atom.constprop.0': > > > > > builtin-kwork.c:(.text+0x5aba8): undefined reference to `work_exit' > > > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `cmd_kwork': > > > > > (.text+0x5bb4b): undefined reference to `work_exit' > > > > > /usr/bin/ld: (.text+0x5bbae): undefined reference to `work_exit' > > > > > /usr/bin/ld: (.text+0x5c3b2): undefined reference to `work_exit' > > > > > collect2: error: ld returned 1 exit status > > > > > make[2]: *** [Makefile.perf:586: /tmp/build/perf/perf] Error 1 > > > > > make[1]: *** [Makefile.perf:289: sub-make] Error 2 > > > > > make: *** [Makefile:76: all] Error 2 > > > > > make: Leaving directory '/git/perf-7.1.0-rc2/tools/perf' > > > > > + exit 1 > > > > > toolsbuilder@five:~$ > > > > > > > > > > > > > > > I'm doing some more tests, just a moment and I'll push what I have to > > > > > tmp.perf-tools-next. > > > > > > > > Also: > > > > > > > > toolsbuilder@five:~$ cat dm.log/almalinux\:8 > > > > <SNIP> > > > > OLLECT_GCC=gcc > > > > COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/8/lto-wrapper > > > > OFFLOAD_TARGET_NAMES=nvptx-none > > > > OFFLOAD_TARGET_DEFAULT=1 > > > > Target: x86_64-redhat-linux > > > > Configured with: ../configure --enable-bootstrap --enable-languages=c,c++,fortran,lto --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugs.almalinux.org/ --enable-shared --enable-threads=posix --enable-checking=release --enable-multilib --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-gcc-major-version-only --with-linker-hash-style=gnu --enable-plugin --enable-initfini-array --with-isl --disable-libmpx --enable-offload-targets=nvptx-none --without-cuda-driver --enable-gnu-indirect-function --enable-cet --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux > > > > Thread model: posix > > > > gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > > > > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= -C tools/perf O=/tmp/build/perf > > > > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' > > > > BUILD: Doing 'make -j32' parallel build > > > > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev > > > > Makefile.config:1061: No libcapstone found, disables disasm engine support for 'perf script', please install libcapstone-dev/capstone-devel > > > > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev > > > > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. > > > > > > > > Auto-detecting system features: > > > > ... libdw: [ on ] > > > > ... glibc: [ on ] > > > > ... libelf: [ on ] > > > > ... libnuma: [ on ] > > > > ... numa_num_possible_cpus: [ on ] > > > > ... libpython: [ on ] > > > > ... libcapstone: [ OFF ] > > > > ... llvm-perf: [ on ] > > > > ... zlib: [ on ] > > > > ... lzma: [ on ] > > > > ... bpf: [ on ] > > > > ... libaio: [ on ] > > > > ... libzstd: [ on ] > > > > ... libopenssl: [ on ] > > > > ... rust: [ OFF ] > > > > > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v2.o > > > > <SNIP> > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/cache.json > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/counter.json > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/floating-point.json > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/frontend.json > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/memory.json > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/other.json > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/pipeline.json > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/virtual-memory.json > > > > GEN /tmp/build/perf/pmu-events/test-empty-pmu-events.c > > > > Traceback (most recent call last): > > > > File "pmu-events/jevents.py", line 1524, in <module> > > > > main() > > > > File "pmu-events/jevents.py", line 1487, in main > > > > with concurrent.futures.ProcessPoolExecutor(initializer=_init_worker, initargs=(_arch_std_events,)) as executor: > > > > TypeError: __init__() got an unexpected keyword argument 'initializer' > > > > make[3]: *** [pmu-events/Build:176: /tmp/build/perf/pmu-events/test-empty-pmu-events.c] Error 1 > > > > make[3]: *** Deleting file '/tmp/build/perf/pmu-events/test-empty-pmu-events.c' > > > > make[2]: *** [Makefile.perf:554: /tmp/build/perf/pmu-events/pmu-events-in.o] Error 2 > > > > make[2]: *** Waiting for unfinished jobs.... > > > > > > But: > > > > > > toolsbuilder@five:~$ export BUILD_TARBALL=http://192.168.86.5/perf/perf-7.1.0-rc2.tar.xz > > > toolsbuilder@five:~$ time dm > > > 1 8.77 almalinux:8 : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > > > 2 84.96 almalinux:9 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-11) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-3.el9) flex 2.6.4 > > > 3: 103.23 almalinux:9-i386 : Ok gcc (GCC) 11.4.1 20231218 (Red Hat 11.4.1-3) , clang version 17.0.6 (AlmaLinux OS Foundation 17.0.6-5.el9) flex 2.6.4 > > > 4 90.91 almalinux:10 : Ok gcc (GCC) 14.3.1 20250617 (Red Hat 14.3.1-2) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-1.el10.alma.1) flex 2.6.4 > > > 5 118.21 alpine:3.16 : Ok gcc (Alpine 11.2.1_git20220219) 11.2.1 20220219 , Alpine clang version 13.0.1 flex 2.6.4 > > > ... > > > > You can drop the parallelize jevent.py changes and I'll follow up. > > Sorry for the breakage. > > It was just that almalinux:8 so far, its an old system, so I'm keeping > it to see what other systems fail, if any. > > So far: > > toolsbuilder@five:~$ export BUILD_TARBALL=http://192.168.86.5/perf/perf-7.1.0-rc2.tar.xz > toolsbuilder@five:~$ time dm > 1 8.77 almalinux:8 : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > 2 84.96 almalinux:9 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-11) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-3.el9) flex 2.6.4 > 3: almalinux:9-i386WARNING: image platform (linux/386) does not match the expected platform (linux/amd64) > WARNING: image platform (linux/386) does not match the expected platform (linux/amd64) > 103.23 almalinux:9-i386 : Ok gcc (GCC) 11.4.1 20231218 (Red Hat 11.4.1-3) , clang version 17.0.6 (AlmaLinux OS Foundation 17.0.6-5.el9) flex 2.6.4 > 4 90.91 almalinux:10 : Ok gcc (GCC) 14.3.1 20250617 (Red Hat 14.3.1-2) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-1.el10.alma.1) flex 2.6.4 > 5 118.21 alpine:3.16 : Ok gcc (Alpine 11.2.1_git20220219) 11.2.1 20220219 , Alpine clang version 13.0.1 flex 2.6.4 > 6 106.13 alpine:3.17 : Ok gcc (Alpine 12.2.1_git20220924-r4) 12.2.1 20220924 , Alpine clang version 15.0.7 flex 2.6.4 > 7 102.80 alpine:3.18 : Ok gcc (Alpine 12.2.1_git20220924-r10) 12.2.1 20220924 , Alpine clang version 16.0.6 flex 2.6.4 > 8 114.48 alpine:3.19 : Ok gcc (Alpine 13.2.1_git20231014) 13.2.1 20231014 , Alpine clang version 17.0.5 flex 2.6.4 > 9 109.53 alpine:3.20 : Ok gcc (Alpine 13.2.1_git20240309) 13.2.1 20240309 , Alpine clang version 17.0.6 flex 2.6.4 > 10 116.20 alpine:3.22 : Ok gcc (Alpine 14.2.0) 14.2.0 , Alpine clang version 20.1.8 flex 2.6.4 > 11 109.83 alpine:3.23 : Ok gcc (Alpine 15.2.0) 15.2.0 , Alpine clang version 21.1.2 flex 2.6.4 > 12 119.23 alpine:edge : Ok gcc (Alpine 14.2.0) 14.2.0 , Alpine clang version 19.1.4 flex 2.6.4 > 13 93.22 amazonlinux:2023 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-5) , clang version 15.0.7 (AWS 15.0.7-3.amzn2023.0.4) flex 2.6.4 > 14 93.40 amazonlinux:devel : Ok gcc (GCC) 11.3.1 20221121 (Red Hat 11.3.1-4) , clang version 15.0.6 (Amazon Linux 15.0.6-3.amzn2023.0.2) flex 2.6.4 > 15 102.98 archlinux:base : Ok gcc (GCC) 15.2.1 20250813 , clang version 20.1.8 flex 2.6.4 > 16 9.58 centos:stream : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-21) (GCC) > 17: clearlinux:latest > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/virtual-memory.json > GEN /tmp/build/perf/pmu-events/test-empty-pmu-events.c > Traceback (most recent call last): > File "pmu-events/jevents.py", line 1524, in <module> > main() > File "pmu-events/jevents.py", line 1487, in main > with concurrent.futures.ProcessPoolExecutor(initializer=_init_worker, initargs=(_arch_std_events,)) as executor: > TypeError: __init__() got an unexpected keyword argument 'initializer' > make[3]: *** [pmu-events/Build:176: /tmp/build/perf/pmu-events/test-empty-pmu-events.c] Error 1 > make[3]: *** Deleting file '/tmp/build/perf/pmu-events/test-empty-pmu-events.c' > make[2]: *** [Makefile.perf:554: /tmp/build/perf/pmu-events/pmu-events-in.o] Error 2 > make[2]: *** Waiting for unfinished jobs.... > > Auto-detecting system features: > ... clang-bpf-co-re: [ on ] > ... llvm: [ on ] > ... libcap: [ on ] > ... libbfd: [ on ] > > MKDIR /tmp/build/perf/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf > > > centos:stream is the bleeding edge, so I'll drop that as you suggested, > ok. So I removed these, i.e. from the parallelize onwards: 28 O T May 18 Ian Rogers ( 43) │ ├─>[PATCH v7 14/14] perf build: Convert llvm-config shell queries to simply expanded variables 29 O T May 18 Ian Rogers ( 62) │ ├─>[PATCH v7 13/14] perf pmu-events: Convert recursive shell assignments and macros to Make built-ins 30 O T May 18 Ian Rogers ( 47) │ ├─>[PATCH v7 12/14] perf build: Prefix SCRIPTS with output directory to fix continuous rebuilds 31 O T May 18 Ian Rogers ( 97) │ ├─>[PATCH v7 11/14] perf pmu-events: Parallelize JSON and metric pre-computation in jevents.py
On Wed, May 20, 2026 at 2:15 PM Arnaldo Carvalho de Melo <acme@kernel.org> wrote: > > On Wed, May 20, 2026 at 06:13:33PM -0300, Arnaldo Carvalho de Melo wrote: > > On Wed, May 20, 2026 at 02:01:03PM -0700, Ian Rogers wrote: > > > On Wed, May 20, 2026 at 1:55 PM Arnaldo Carvalho de Melo > > > <acme@kernel.org> wrote: > > > > > > > > On Wed, May 20, 2026 at 05:51:36PM -0300, Arnaldo Carvalho de Melo wrote: > > > > > On Wed, May 20, 2026 at 05:49:57PM -0300, Arnaldo Carvalho de Melo wrote: > > > > > > On Wed, May 20, 2026 at 01:40:48PM -0700, Ian Rogers wrote: > > > > > > > On Wed, May 20, 2026 at 12:46 PM Arnaldo Carvalho de Melo > > > > > > > <acme@kernel.org> wrote: > > > > > > > > > > > > > > > > On Wed, May 20, 2026 at 04:15:30PM -0300, Arnaldo Carvalho de Melo wrote: > > > > > > > > > On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote: > > > > > > > > > > Nearly all perf code ends up passing an evsel with the perf_sample, > > > > > > > > > > which is problematic if you want to rewrite the evsel (such as with > > > > > > > > > > off-CPU processing) because all uses of the evsel need to be fixed > > > > > > > > > > up. Since the perf_sample now carries its own resolved evsel pointer, > > > > > > > > > > passing evsel separately is redundant and error-prone (allowing > > > > > > > > > > parameter divergence bugs). > > > > > > > > > > > > > > > > > > > > This series cleans up the redundant evsel parameter across all perf > > > > > > > > > > tool subcommands and APIs, ensuring that tools uniformly retrieve > > > > > > > > > > event attributes directly from the sample itself. This simplifies > > > > > > > > > > function signatures and improves API consistency. > > > > > > > > > > > > > > > > > > > > Additionally, this series incorporates subsequent memory safety > > > > > > > > > > improvements, robust bounds checking, and memory leak corrections > > > > > > > > > > prompted by address sanitizers and Sashiko code reviews. > > > > > > > > > > > > > > > > > > > > The original combined RFC patch: > > > > > > > > > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > > > > > > > > > > was split up per reviewer requests. > > > > > > > > > > > > > > > > > > > > v14: > > > > > > > > > > - Incorporate Acked-by tags from Namhyung Kim. > > > > > > > > > > - Re-polish and expand commit descriptions to provide thorough explanations > > > > > > > > > > of the changes and their rationale. > > > > > > > > > > > > > > > > > > Thanks, tentatively applying it to perf-tools-next, for v7.2. > > > > > > > > > > > > > > > > Had to apply a minor fixup for a conflict with anoter patch series from > > > > > > > > you, the LBR one. > > > > > > > > > > > > > > Sorry for that and thanks for dealing with the conflict! Any chance > > > > > > > you could push to tmp.perf-tools-next? I have some other series that > > > > > > > will conflict and it would be nice to rebase those changes there. > > > > > > > > > > > > I just removed the last two patches, the ones for kwork, as: > > > > > > > > > > > > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= NO_LIBELF=1 -C tools/perf O=/tmp/build/perf > > > > > > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' > > > > > > BUILD: Doing 'make -j32' parallel build > > > > > > Makefile.config:647: Warning: Disabled BPF skeletons as libelf is required by bpftool > > > > > > Makefile.config:745: Disabling post unwind, no support found. > > > > > > Makefile.config:856: Missing python setuptools, the python binding won't be built, please install python3-setuptools or equivalent > > > > > > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev > > > > > > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev > > > > > > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. > > > > > > > > > > > > Auto-detecting system features: > > > > > > ... libdw: [ on ] > > > > > > ... glibc: [ on ] > > > > > > ... libelf: [ on ] > > > > > > ... libnuma: [ on ] > > > > > > ... numa_num_possible_cpus: [ on ] > > > > > > ... libpython: [ on ] > > > > > > ... libcapstone: [ on ] > > > > > > ... llvm-perf: [ on ] > > > > > > ... zlib: [ on ] > > > > > > ... lzma: [ on ] > > > > > > ... bpf: [ on ] > > > > > > ... libaio: [ on ] > > > > > > ... libzstd: [ on ] > > > > > > ... libopenssl: [ on ] > > > > > > ... rust: [ OFF ] > > > > > > > > > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o > > > > > > <SNIP> > > > > > > LINK /tmp/build/perf/perf > > > > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_pop_atom': > > > > > > builtin-kwork.c:(.text+0x59c92): undefined reference to `work_exit' > > > > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_push_atom.constprop.0': > > > > > > builtin-kwork.c:(.text+0x5aba8): undefined reference to `work_exit' > > > > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `cmd_kwork': > > > > > > (.text+0x5bb4b): undefined reference to `work_exit' > > > > > > /usr/bin/ld: (.text+0x5bbae): undefined reference to `work_exit' > > > > > > /usr/bin/ld: (.text+0x5c3b2): undefined reference to `work_exit' > > > > > > collect2: error: ld returned 1 exit status > > > > > > make[2]: *** [Makefile.perf:586: /tmp/build/perf/perf] Error 1 > > > > > > make[1]: *** [Makefile.perf:289: sub-make] Error 2 > > > > > > make: *** [Makefile:76: all] Error 2 > > > > > > make: Leaving directory '/git/perf-7.1.0-rc2/tools/perf' > > > > > > + exit 1 > > > > > > toolsbuilder@five:~$ > > > > > > > > > > > > > > > > > > I'm doing some more tests, just a moment and I'll push what I have to > > > > > > tmp.perf-tools-next. > > > > > > > > > > Also: > > > > > > > > > > toolsbuilder@five:~$ cat dm.log/almalinux\:8 > > > > > <SNIP> > > > > > OLLECT_GCC=gcc > > > > > COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/8/lto-wrapper > > > > > OFFLOAD_TARGET_NAMES=nvptx-none > > > > > OFFLOAD_TARGET_DEFAULT=1 > > > > > Target: x86_64-redhat-linux > > > > > Configured with: ../configure --enable-bootstrap --enable-languages=c,c++,fortran,lto --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugs.almalinux.org/ --enable-shared --enable-threads=posix --enable-checking=release --enable-multilib --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-gcc-major-version-only --with-linker-hash-style=gnu --enable-plugin --enable-initfini-array --with-isl --disable-libmpx --enable-offload-targets=nvptx-none --without-cuda-driver --enable-gnu-indirect-function --enable-cet --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux > > > > > Thread model: posix > > > > > gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > > > > > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= -C tools/perf O=/tmp/build/perf > > > > > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' > > > > > BUILD: Doing 'make -j32' parallel build > > > > > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev > > > > > Makefile.config:1061: No libcapstone found, disables disasm engine support for 'perf script', please install libcapstone-dev/capstone-devel > > > > > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev > > > > > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. > > > > > > > > > > Auto-detecting system features: > > > > > ... libdw: [ on ] > > > > > ... glibc: [ on ] > > > > > ... libelf: [ on ] > > > > > ... libnuma: [ on ] > > > > > ... numa_num_possible_cpus: [ on ] > > > > > ... libpython: [ on ] > > > > > ... libcapstone: [ OFF ] > > > > > ... llvm-perf: [ on ] > > > > > ... zlib: [ on ] > > > > > ... lzma: [ on ] > > > > > ... bpf: [ on ] > > > > > ... libaio: [ on ] > > > > > ... libzstd: [ on ] > > > > > ... libopenssl: [ on ] > > > > > ... rust: [ OFF ] > > > > > > > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o > > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v2.o > > > > > <SNIP> > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/cache.json > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/counter.json > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/floating-point.json > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/frontend.json > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/memory.json > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/other.json > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/pipeline.json > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/virtual-memory.json > > > > > GEN /tmp/build/perf/pmu-events/test-empty-pmu-events.c > > > > > Traceback (most recent call last): > > > > > File "pmu-events/jevents.py", line 1524, in <module> > > > > > main() > > > > > File "pmu-events/jevents.py", line 1487, in main > > > > > with concurrent.futures.ProcessPoolExecutor(initializer=_init_worker, initargs=(_arch_std_events,)) as executor: > > > > > TypeError: __init__() got an unexpected keyword argument 'initializer' > > > > > make[3]: *** [pmu-events/Build:176: /tmp/build/perf/pmu-events/test-empty-pmu-events.c] Error 1 > > > > > make[3]: *** Deleting file '/tmp/build/perf/pmu-events/test-empty-pmu-events.c' > > > > > make[2]: *** [Makefile.perf:554: /tmp/build/perf/pmu-events/pmu-events-in.o] Error 2 > > > > > make[2]: *** Waiting for unfinished jobs.... > > > > > > > > But: > > > > > > > > toolsbuilder@five:~$ export BUILD_TARBALL=http://192.168.86.5/perf/perf-7.1.0-rc2.tar.xz > > > > toolsbuilder@five:~$ time dm > > > > 1 8.77 almalinux:8 : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > > > > 2 84.96 almalinux:9 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-11) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-3.el9) flex 2.6.4 > > > > 3: 103.23 almalinux:9-i386 : Ok gcc (GCC) 11.4.1 20231218 (Red Hat 11.4.1-3) , clang version 17.0.6 (AlmaLinux OS Foundation 17.0.6-5.el9) flex 2.6.4 > > > > 4 90.91 almalinux:10 : Ok gcc (GCC) 14.3.1 20250617 (Red Hat 14.3.1-2) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-1.el10.alma.1) flex 2.6.4 > > > > 5 118.21 alpine:3.16 : Ok gcc (Alpine 11.2.1_git20220219) 11.2.1 20220219 , Alpine clang version 13.0.1 flex 2.6.4 > > > > ... > > > > > > You can drop the parallelize jevent.py changes and I'll follow up. > > > Sorry for the breakage. > > > > It was just that almalinux:8 so far, its an old system, so I'm keeping > > it to see what other systems fail, if any. > > > > So far: > > > > toolsbuilder@five:~$ export BUILD_TARBALL=http://192.168.86.5/perf/perf-7.1.0-rc2.tar.xz > > toolsbuilder@five:~$ time dm > > 1 8.77 almalinux:8 : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > > 2 84.96 almalinux:9 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-11) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-3.el9) flex 2.6.4 > > 3: almalinux:9-i386WARNING: image platform (linux/386) does not match the expected platform (linux/amd64) > > WARNING: image platform (linux/386) does not match the expected platform (linux/amd64) > > 103.23 almalinux:9-i386 : Ok gcc (GCC) 11.4.1 20231218 (Red Hat 11.4.1-3) , clang version 17.0.6 (AlmaLinux OS Foundation 17.0.6-5.el9) flex 2.6.4 > > 4 90.91 almalinux:10 : Ok gcc (GCC) 14.3.1 20250617 (Red Hat 14.3.1-2) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-1.el10.alma.1) flex 2.6.4 > > 5 118.21 alpine:3.16 : Ok gcc (Alpine 11.2.1_git20220219) 11.2.1 20220219 , Alpine clang version 13.0.1 flex 2.6.4 > > 6 106.13 alpine:3.17 : Ok gcc (Alpine 12.2.1_git20220924-r4) 12.2.1 20220924 , Alpine clang version 15.0.7 flex 2.6.4 > > 7 102.80 alpine:3.18 : Ok gcc (Alpine 12.2.1_git20220924-r10) 12.2.1 20220924 , Alpine clang version 16.0.6 flex 2.6.4 > > 8 114.48 alpine:3.19 : Ok gcc (Alpine 13.2.1_git20231014) 13.2.1 20231014 , Alpine clang version 17.0.5 flex 2.6.4 > > 9 109.53 alpine:3.20 : Ok gcc (Alpine 13.2.1_git20240309) 13.2.1 20240309 , Alpine clang version 17.0.6 flex 2.6.4 > > 10 116.20 alpine:3.22 : Ok gcc (Alpine 14.2.0) 14.2.0 , Alpine clang version 20.1.8 flex 2.6.4 > > 11 109.83 alpine:3.23 : Ok gcc (Alpine 15.2.0) 15.2.0 , Alpine clang version 21.1.2 flex 2.6.4 > > 12 119.23 alpine:edge : Ok gcc (Alpine 14.2.0) 14.2.0 , Alpine clang version 19.1.4 flex 2.6.4 > > 13 93.22 amazonlinux:2023 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-5) , clang version 15.0.7 (AWS 15.0.7-3.amzn2023.0.4) flex 2.6.4 > > 14 93.40 amazonlinux:devel : Ok gcc (GCC) 11.3.1 20221121 (Red Hat 11.3.1-4) , clang version 15.0.6 (Amazon Linux 15.0.6-3.amzn2023.0.2) flex 2.6.4 > > 15 102.98 archlinux:base : Ok gcc (GCC) 15.2.1 20250813 , clang version 20.1.8 flex 2.6.4 > > 16 9.58 centos:stream : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-21) (GCC) > > 17: clearlinux:latest > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/virtual-memory.json > > GEN /tmp/build/perf/pmu-events/test-empty-pmu-events.c > > Traceback (most recent call last): > > File "pmu-events/jevents.py", line 1524, in <module> > > main() > > File "pmu-events/jevents.py", line 1487, in main > > with concurrent.futures.ProcessPoolExecutor(initializer=_init_worker, initargs=(_arch_std_events,)) as executor: > > TypeError: __init__() got an unexpected keyword argument 'initializer' > > make[3]: *** [pmu-events/Build:176: /tmp/build/perf/pmu-events/test-empty-pmu-events.c] Error 1 > > make[3]: *** Deleting file '/tmp/build/perf/pmu-events/test-empty-pmu-events.c' > > make[2]: *** [Makefile.perf:554: /tmp/build/perf/pmu-events/pmu-events-in.o] Error 2 > > make[2]: *** Waiting for unfinished jobs.... > > > > Auto-detecting system features: > > ... clang-bpf-co-re: [ on ] > > ... llvm: [ on ] > > ... libcap: [ on ] > > ... libbfd: [ on ] > > > > MKDIR /tmp/build/perf/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf > > > > > > centos:stream is the bleeding edge, so I'll drop that as you suggested, > > ok. Sgtm. I'm confused though, something old and something new broke? In the "Minimal requirements to compile the Kernel" (https://www.kernel.org/doc/html/latest/process/changes.html) they state Python must be version 3.9, so it'd be nice to set that as the baseline wrt APIs. We could probably automate that in the build and make the build NO_PYTHON if the version is older than 3.9. > So I removed these, i.e. from the parallelize onwards: > > 28 O T May 18 Ian Rogers ( 43) │ ├─>[PATCH v7 14/14] perf build: Convert llvm-config shell queries to simply expanded variables > 29 O T May 18 Ian Rogers ( 62) │ ├─>[PATCH v7 13/14] perf pmu-events: Convert recursive shell assignments and macros to Make built-ins > 30 O T May 18 Ian Rogers ( 47) │ ├─>[PATCH v7 12/14] perf build: Prefix SCRIPTS with output directory to fix continuous rebuilds > 31 O T May 18 Ian Rogers ( 97) │ ├─>[PATCH v7 11/14] perf pmu-events: Parallelize JSON and metric pre-computation in jevents.py > Could we keep 12, 13 and 14? They are independent changes and offer significant build performance improvements. Thanks, Ian
On Wed, May 20, 2026 at 02:18:46PM -0700, Ian Rogers wrote: > On Wed, May 20, 2026 at 2:15 PM Arnaldo Carvalho de Melo > <acme@kernel.org> wrote: > > > > On Wed, May 20, 2026 at 06:13:33PM -0300, Arnaldo Carvalho de Melo wrote: > > > On Wed, May 20, 2026 at 02:01:03PM -0700, Ian Rogers wrote: > > > > On Wed, May 20, 2026 at 1:55 PM Arnaldo Carvalho de Melo > > > > <acme@kernel.org> wrote: > > > > > > > > > > On Wed, May 20, 2026 at 05:51:36PM -0300, Arnaldo Carvalho de Melo wrote: > > > > > > On Wed, May 20, 2026 at 05:49:57PM -0300, Arnaldo Carvalho de Melo wrote: > > > > > > > On Wed, May 20, 2026 at 01:40:48PM -0700, Ian Rogers wrote: > > > > > > > > On Wed, May 20, 2026 at 12:46 PM Arnaldo Carvalho de Melo > > > > > > > > <acme@kernel.org> wrote: > > > > > > > > > > > > > > > > > > On Wed, May 20, 2026 at 04:15:30PM -0300, Arnaldo Carvalho de Melo wrote: > > > > > > > > > > On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote: > > > > > > > > > > > Nearly all perf code ends up passing an evsel with the perf_sample, > > > > > > > > > > > which is problematic if you want to rewrite the evsel (such as with > > > > > > > > > > > off-CPU processing) because all uses of the evsel need to be fixed > > > > > > > > > > > up. Since the perf_sample now carries its own resolved evsel pointer, > > > > > > > > > > > passing evsel separately is redundant and error-prone (allowing > > > > > > > > > > > parameter divergence bugs). > > > > > > > > > > > > > > > > > > > > > > This series cleans up the redundant evsel parameter across all perf > > > > > > > > > > > tool subcommands and APIs, ensuring that tools uniformly retrieve > > > > > > > > > > > event attributes directly from the sample itself. This simplifies > > > > > > > > > > > function signatures and improves API consistency. > > > > > > > > > > > > > > > > > > > > > > Additionally, this series incorporates subsequent memory safety > > > > > > > > > > > improvements, robust bounds checking, and memory leak corrections > > > > > > > > > > > prompted by address sanitizers and Sashiko code reviews. > > > > > > > > > > > > > > > > > > > > > > The original combined RFC patch: > > > > > > > > > > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > > > > > > > > > > > was split up per reviewer requests. > > > > > > > > > > > > > > > > > > > > > > v14: > > > > > > > > > > > - Incorporate Acked-by tags from Namhyung Kim. > > > > > > > > > > > - Re-polish and expand commit descriptions to provide thorough explanations > > > > > > > > > > > of the changes and their rationale. > > > > > > > > > > > > > > > > > > > > Thanks, tentatively applying it to perf-tools-next, for v7.2. > > > > > > > > > > > > > > > > > > Had to apply a minor fixup for a conflict with anoter patch series from > > > > > > > > > you, the LBR one. > > > > > > > > > > > > > > > > Sorry for that and thanks for dealing with the conflict! Any chance > > > > > > > > you could push to tmp.perf-tools-next? I have some other series that > > > > > > > > will conflict and it would be nice to rebase those changes there. > > > > > > > > > > > > > > I just removed the last two patches, the ones for kwork, as: > > > > > > > > > > > > > > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= NO_LIBELF=1 -C tools/perf O=/tmp/build/perf > > > > > > > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' > > > > > > > BUILD: Doing 'make -j32' parallel build > > > > > > > Makefile.config:647: Warning: Disabled BPF skeletons as libelf is required by bpftool > > > > > > > Makefile.config:745: Disabling post unwind, no support found. > > > > > > > Makefile.config:856: Missing python setuptools, the python binding won't be built, please install python3-setuptools or equivalent > > > > > > > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev > > > > > > > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev > > > > > > > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. > > > > > > > > > > > > > > Auto-detecting system features: > > > > > > > ... libdw: [ on ] > > > > > > > ... glibc: [ on ] > > > > > > > ... libelf: [ on ] > > > > > > > ... libnuma: [ on ] > > > > > > > ... numa_num_possible_cpus: [ on ] > > > > > > > ... libpython: [ on ] > > > > > > > ... libcapstone: [ on ] > > > > > > > ... llvm-perf: [ on ] > > > > > > > ... zlib: [ on ] > > > > > > > ... lzma: [ on ] > > > > > > > ... bpf: [ on ] > > > > > > > ... libaio: [ on ] > > > > > > > ... libzstd: [ on ] > > > > > > > ... libopenssl: [ on ] > > > > > > > ... rust: [ OFF ] > > > > > > > > > > > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o > > > > > > > <SNIP> > > > > > > > LINK /tmp/build/perf/perf > > > > > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_pop_atom': > > > > > > > builtin-kwork.c:(.text+0x59c92): undefined reference to `work_exit' > > > > > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `work_push_atom.constprop.0': > > > > > > > builtin-kwork.c:(.text+0x5aba8): undefined reference to `work_exit' > > > > > > > /usr/bin/ld: /tmp/build/perf/perf-in.o: in function `cmd_kwork': > > > > > > > (.text+0x5bb4b): undefined reference to `work_exit' > > > > > > > /usr/bin/ld: (.text+0x5bbae): undefined reference to `work_exit' > > > > > > > /usr/bin/ld: (.text+0x5c3b2): undefined reference to `work_exit' > > > > > > > collect2: error: ld returned 1 exit status > > > > > > > make[2]: *** [Makefile.perf:586: /tmp/build/perf/perf] Error 1 > > > > > > > make[1]: *** [Makefile.perf:289: sub-make] Error 2 > > > > > > > make: *** [Makefile:76: all] Error 2 > > > > > > > make: Leaving directory '/git/perf-7.1.0-rc2/tools/perf' > > > > > > > + exit 1 > > > > > > > toolsbuilder@five:~$ > > > > > > > > > > > > > > > > > > > > > I'm doing some more tests, just a moment and I'll push what I have to > > > > > > > tmp.perf-tools-next. > > > > > > > > > > > > Also: > > > > > > > > > > > > toolsbuilder@five:~$ cat dm.log/almalinux\:8 > > > > > > <SNIP> > > > > > > OLLECT_GCC=gcc > > > > > > COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/8/lto-wrapper > > > > > > OFFLOAD_TARGET_NAMES=nvptx-none > > > > > > OFFLOAD_TARGET_DEFAULT=1 > > > > > > Target: x86_64-redhat-linux > > > > > > Configured with: ../configure --enable-bootstrap --enable-languages=c,c++,fortran,lto --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugs.almalinux.org/ --enable-shared --enable-threads=posix --enable-checking=release --enable-multilib --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-gcc-major-version-only --with-linker-hash-style=gnu --enable-plugin --enable-initfini-array --with-isl --disable-libmpx --enable-offload-targets=nvptx-none --without-cuda-driver --enable-gnu-indirect-function --enable-cet --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux > > > > > > Thread model: posix > > > > > > gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > > > > > > + make ARCH= CROSS_COMPILE= EXTRA_CFLAGS= -C tools/perf O=/tmp/build/perf > > > > > > make: Entering directory '/git/perf-7.1.0-rc2/tools/perf' > > > > > > BUILD: Doing 'make -j32' parallel build > > > > > > Makefile.config:1045: No libbabeltrace found, disables 'perf data' CTF format support, please install libbabeltrace-dev[el]/libbabeltrace-ctf-dev > > > > > > Makefile.config:1061: No libcapstone found, disables disasm engine support for 'perf script', please install libcapstone-dev/capstone-devel > > > > > > Makefile.config:1110: libpfm4 not found, disables libpfm4 support. Please install libpfm-devel or libpfm4-dev > > > > > > Makefile.config:1135: Rust is not found. Test workloads with rust are disabled. > > > > > > > > > > > > Auto-detecting system features: > > > > > > ... libdw: [ on ] > > > > > > ... glibc: [ on ] > > > > > > ... libelf: [ on ] > > > > > > ... libnuma: [ on ] > > > > > > ... numa_num_possible_cpus: [ on ] > > > > > > ... libpython: [ on ] > > > > > > ... libcapstone: [ OFF ] > > > > > > ... llvm-perf: [ on ] > > > > > > ... zlib: [ on ] > > > > > > ... lzma: [ on ] > > > > > > ... bpf: [ on ] > > > > > > ... libaio: [ on ] > > > > > > ... libzstd: [ on ] > > > > > > ... libopenssl: [ on ] > > > > > > ... rust: [ OFF ] > > > > > > > > > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v0.o > > > > > > CC /tmp/build/perf/dlfilters/dlfilter-test-api-v2.o > > > > > > <SNIP> > > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/cache.json > > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/counter.json > > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/floating-point.json > > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/frontend.json > > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/memory.json > > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/other.json > > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/pipeline.json > > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/virtual-memory.json > > > > > > GEN /tmp/build/perf/pmu-events/test-empty-pmu-events.c > > > > > > Traceback (most recent call last): > > > > > > File "pmu-events/jevents.py", line 1524, in <module> > > > > > > main() > > > > > > File "pmu-events/jevents.py", line 1487, in main > > > > > > with concurrent.futures.ProcessPoolExecutor(initializer=_init_worker, initargs=(_arch_std_events,)) as executor: > > > > > > TypeError: __init__() got an unexpected keyword argument 'initializer' > > > > > > make[3]: *** [pmu-events/Build:176: /tmp/build/perf/pmu-events/test-empty-pmu-events.c] Error 1 > > > > > > make[3]: *** Deleting file '/tmp/build/perf/pmu-events/test-empty-pmu-events.c' > > > > > > make[2]: *** [Makefile.perf:554: /tmp/build/perf/pmu-events/pmu-events-in.o] Error 2 > > > > > > make[2]: *** Waiting for unfinished jobs.... > > > > > > > > > > But: > > > > > > > > > > toolsbuilder@five:~$ export BUILD_TARBALL=http://192.168.86.5/perf/perf-7.1.0-rc2.tar.xz > > > > > toolsbuilder@five:~$ time dm > > > > > 1 8.77 almalinux:8 : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > > > > > 2 84.96 almalinux:9 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-11) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-3.el9) flex 2.6.4 > > > > > 3: 103.23 almalinux:9-i386 : Ok gcc (GCC) 11.4.1 20231218 (Red Hat 11.4.1-3) , clang version 17.0.6 (AlmaLinux OS Foundation 17.0.6-5.el9) flex 2.6.4 > > > > > 4 90.91 almalinux:10 : Ok gcc (GCC) 14.3.1 20250617 (Red Hat 14.3.1-2) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-1.el10.alma.1) flex 2.6.4 > > > > > 5 118.21 alpine:3.16 : Ok gcc (Alpine 11.2.1_git20220219) 11.2.1 20220219 , Alpine clang version 13.0.1 flex 2.6.4 > > > > > ... > > > > > > > > You can drop the parallelize jevent.py changes and I'll follow up. > > > > Sorry for the breakage. > > > > > > It was just that almalinux:8 so far, its an old system, so I'm keeping > > > it to see what other systems fail, if any. > > > > > > So far: > > > > > > toolsbuilder@five:~$ export BUILD_TARBALL=http://192.168.86.5/perf/perf-7.1.0-rc2.tar.xz > > > toolsbuilder@five:~$ time dm > > > 1 8.77 almalinux:8 : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-28) (GCC) > > > 2 84.96 almalinux:9 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-11) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-3.el9) flex 2.6.4 > > > 3: almalinux:9-i386WARNING: image platform (linux/386) does not match the expected platform (linux/amd64) > > > WARNING: image platform (linux/386) does not match the expected platform (linux/amd64) > > > 103.23 almalinux:9-i386 : Ok gcc (GCC) 11.4.1 20231218 (Red Hat 11.4.1-3) , clang version 17.0.6 (AlmaLinux OS Foundation 17.0.6-5.el9) flex 2.6.4 > > > 4 90.91 almalinux:10 : Ok gcc (GCC) 14.3.1 20250617 (Red Hat 14.3.1-2) , clang version 20.1.8 (AlmaLinux OS Foundation 20.1.8-1.el10.alma.1) flex 2.6.4 > > > 5 118.21 alpine:3.16 : Ok gcc (Alpine 11.2.1_git20220219) 11.2.1 20220219 , Alpine clang version 13.0.1 flex 2.6.4 > > > 6 106.13 alpine:3.17 : Ok gcc (Alpine 12.2.1_git20220924-r4) 12.2.1 20220924 , Alpine clang version 15.0.7 flex 2.6.4 > > > 7 102.80 alpine:3.18 : Ok gcc (Alpine 12.2.1_git20220924-r10) 12.2.1 20220924 , Alpine clang version 16.0.6 flex 2.6.4 > > > 8 114.48 alpine:3.19 : Ok gcc (Alpine 13.2.1_git20231014) 13.2.1 20231014 , Alpine clang version 17.0.5 flex 2.6.4 > > > 9 109.53 alpine:3.20 : Ok gcc (Alpine 13.2.1_git20240309) 13.2.1 20240309 , Alpine clang version 17.0.6 flex 2.6.4 > > > 10 116.20 alpine:3.22 : Ok gcc (Alpine 14.2.0) 14.2.0 , Alpine clang version 20.1.8 flex 2.6.4 > > > 11 109.83 alpine:3.23 : Ok gcc (Alpine 15.2.0) 15.2.0 , Alpine clang version 21.1.2 flex 2.6.4 > > > 12 119.23 alpine:edge : Ok gcc (Alpine 14.2.0) 14.2.0 , Alpine clang version 19.1.4 flex 2.6.4 > > > 13 93.22 amazonlinux:2023 : Ok gcc (GCC) 11.5.0 20240719 (Red Hat 11.5.0-5) , clang version 15.0.7 (AWS 15.0.7-3.amzn2023.0.4) flex 2.6.4 > > > 14 93.40 amazonlinux:devel : Ok gcc (GCC) 11.3.1 20221121 (Red Hat 11.3.1-4) , clang version 15.0.6 (Amazon Linux 15.0.6-3.amzn2023.0.2) flex 2.6.4 > > > 15 102.98 archlinux:base : Ok gcc (GCC) 15.2.1 20250813 , clang version 20.1.8 flex 2.6.4 > > > 16 9.58 centos:stream : FAIL gcc version 8.5.0 20210514 (Red Hat 8.5.0-21) (GCC) > > > 17: clearlinux:latest > > > > > > COPY /tmp/build/perf/ <- pmu-events/arch/x86/westmereex/virtual-memory.json > > > GEN /tmp/build/perf/pmu-events/test-empty-pmu-events.c > > > Traceback (most recent call last): > > > File "pmu-events/jevents.py", line 1524, in <module> > > > main() > > > File "pmu-events/jevents.py", line 1487, in main > > > with concurrent.futures.ProcessPoolExecutor(initializer=_init_worker, initargs=(_arch_std_events,)) as executor: > > > TypeError: __init__() got an unexpected keyword argument 'initializer' > > > make[3]: *** [pmu-events/Build:176: /tmp/build/perf/pmu-events/test-empty-pmu-events.c] Error 1 > > > make[3]: *** Deleting file '/tmp/build/perf/pmu-events/test-empty-pmu-events.c' > > > make[2]: *** [Makefile.perf:554: /tmp/build/perf/pmu-events/pmu-events-in.o] Error 2 > > > make[2]: *** Waiting for unfinished jobs.... > > > > > > Auto-detecting system features: > > > ... clang-bpf-co-re: [ on ] > > > ... llvm: [ on ] > > > ... libcap: [ on ] > > > ... libbfd: [ on ] > > > > > > MKDIR /tmp/build/perf/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf > > > > > > > > > centos:stream is the bleeding edge, so I'll drop that as you suggested, > > > ok. > > Sgtm. I'm confused though, something old and something new broke? > In the "Minimal requirements to compile the Kernel" > (https://www.kernel.org/doc/html/latest/process/changes.html) they > state Python must be version 3.9, so it'd be nice to set that as the > baseline wrt APIs. We could probably automate that in the build and > make the build NO_PYTHON if the version is older than 3.9. > > > So I removed these, i.e. from the parallelize onwards: > > > > 28 O T May 18 Ian Rogers ( 43) │ ├─>[PATCH v7 14/14] perf build: Convert llvm-config shell queries to simply expanded variables > > 29 O T May 18 Ian Rogers ( 62) │ ├─>[PATCH v7 13/14] perf pmu-events: Convert recursive shell assignments and macros to Make built-ins > > 30 O T May 18 Ian Rogers ( 47) │ ├─>[PATCH v7 12/14] perf build: Prefix SCRIPTS with output directory to fix continuous rebuilds > > 31 O T May 18 Ian Rogers ( 97) │ ├─>[PATCH v7 11/14] perf pmu-events: Parallelize JSON and metric pre-computation in jevents.py > > > > Could we keep 12, 13 and 14? They are independent changes and offer > significant build performance improvements. Lemme try cherry picking them then. I removed without checking if there were any deps. - Arnaldo
Everything is at https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git tmp.perf-tools-next I also rebased my perf.data validation (that has some extra fixes, because... Sashiko): https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git perf-data-validation I think I addressed some of your review comments, if bored, take a look, I'll submit as soon as I get what is in tmp.perf-tools-next moved to perf-tools-next which I plan/hope to do later today, after the container build tests finish. - Arnaldo
On Wed, May 20, 2026 at 2:51 PM Arnaldo Carvalho de Melo <acme@kernel.org> wrote: > > > Everything is at > > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git tmp.perf-tools-next > > I also rebased my perf.data validation (that has some extra fixes, because... Sashiko): > > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git perf-data-validation > > I think I addressed some of your review comments, if bored, take a look, > I'll submit as soon as I get what is in tmp.perf-tools-next moved to > perf-tools-next which I plan/hope to do later today, after the container > build tests finish. I'll look at the perf-data-validation branch. Would it be okay to post the patches to LKML so that we can get the Sashiko review on them? Thanks, Ian > - Arnaldo
On Wed, May 20, 2026 at 03:32:18PM -0700, Ian Rogers wrote: > On Wed, May 20, 2026 at 2:51 PM Arnaldo Carvalho de Melo > <acme@kernel.org> wrote: > > > > > > Everything is at > > > > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git tmp.perf-tools-next > > > > I also rebased my perf.data validation (that has some extra fixes, because... Sashiko): > > > > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git perf-data-validation > > > > I think I addressed some of your review comments, if bored, take a look, > > I'll submit as soon as I get what is in tmp.perf-tools-next moved to > > perf-tools-next which I plan/hope to do later today, after the container > > build tests finish. > > I'll look at the perf-data-validation branch. Would it be okay to post > the patches to LKML so that we can get the Sashiko review on them? I did it, and having passed thru sashiko locally dozens of times it ended up being rather clean :-) I'm addressing some of the valid concerns raised now and will send a v2. - Arnaldo
On Wed, May 20, 2026 at 3:32 PM Ian Rogers <irogers@google.com> wrote: > > On Wed, May 20, 2026 at 2:51 PM Arnaldo Carvalho de Melo > <acme@kernel.org> wrote: > > > > > > Everything is at > > > > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git tmp.perf-tools-next > > > > I also rebased my perf.data validation (that has some extra fixes, because... Sashiko): > > > > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git perf-data-validation > > > > I think I addressed some of your review comments, if bored, take a look, > > I'll submit as soon as I get what is in tmp.perf-tools-next moved to > > perf-tools-next which I plan/hope to do later today, after the container > > build tests finish. > > I'll look at the perf-data-validation branch. Would it be okay to post > the patches to LKML so that we can get the Sashiko review on them? In "perf session: Bound nr_cpus_avail and validate sample CPU" the perf_session__deliver_event change appears to inline evlist__parse_sample which seems unfortunate from a code reuse point-of-view. The commit message calls this out but doesn't describe the motivation for the change. Could we fix evlist__parse_sample? Thanks, Ian > Thanks, > Ian > > > - Arnaldo
On Wed, May 20, 2026 at 03:43:52PM -0700, Ian Rogers wrote:
> On Wed, May 20, 2026 at 3:32 PM Ian Rogers <irogers@google.com> wrote:
> > I'll look at the perf-data-validation branch. Would it be okay to post
> > the patches to LKML so that we can get the Sashiko review on them?
> In "perf session: Bound nr_cpus_avail and validate sample CPU" the
> perf_session__deliver_event change appears to inline
> evlist__parse_sample which seems unfortunate from a code reuse
> point-of-view. The commit message calls this out but doesn't describe
> the motivation for the change. Could we fix evlist__parse_sample?
The motivation:
18 -Also refactor the sample parsing in perf_session__deliver_event()
19 -to call evsel__parse_sample() directly (via evlist__event2evsel()
20 -for evsel lookup), with explicit guest VM SID resolution for
21 -machine_pid and vcpu fields.
18 +Inline evlist__parse_sample() into perf_session__deliver_event()
19 +so the evsel lookup needed for sample_type checking reuses the same
20 +evsel that parsed the sample, avoiding a second evlist__event2evsel()
21 +call on every event.
A follow up patch could introduce a evlist__parse_sample variation that
passes a evsel pointer address where to store the evsel used so that it
can be reused and avoid the duplicate lookup.
- Arnaldo
On Wed, May 20, 2026 at 12:15 PM Arnaldo Carvalho de Melo <acme@kernel.org> wrote: > > On Wed, May 20, 2026 at 12:05:06PM -0700, Ian Rogers wrote: > > Nearly all perf code ends up passing an evsel with the perf_sample, > > which is problematic if you want to rewrite the evsel (such as with > > off-CPU processing) because all uses of the evsel need to be fixed > > up. Since the perf_sample now carries its own resolved evsel pointer, > > passing evsel separately is redundant and error-prone (allowing > > parameter divergence bugs). > > > > This series cleans up the redundant evsel parameter across all perf > > tool subcommands and APIs, ensuring that tools uniformly retrieve > > event attributes directly from the sample itself. This simplifies > > function signatures and improves API consistency. > > > > Additionally, this series incorporates subsequent memory safety > > improvements, robust bounds checking, and memory leak corrections > > prompted by address sanitizers and Sashiko code reviews. > > > > The original combined RFC patch: > > https://lore.kernel.org/lkml/20260126071822.447368-1-irogers@google.com/ > > was split up per reviewer requests. > > > > v14: > > - Incorporate Acked-by tags from Namhyung Kim. > > - Re-polish and expand commit descriptions to provide thorough explanations > > of the changes and their rationale. > > Thanks, tentatively applying it to perf-tools-next, for v7.2. > > Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Thanks Arnaldo! There will be Sashiko reviews to come and the boundary checking is likely to conflict with the boundary checking you are adding in: https://lore.kernel.org/linux-perf-users/20260510033424.255812-1-acme@kernel.org/ When feedback arrives, I prefer sending additional fixes rather than issuing a v15. I don't mind if you want to drop the boundary checks I added, they were only included in the series to satisfy Sashiko and because I worried we'd forget about them. Thanks, Ian
As struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from tool-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-annotate.c | 3 +--
tools/perf/builtin-c2c.c | 2 +-
tools/perf/builtin-diff.c | 2 +-
tools/perf/builtin-inject.c | 19 ++++++---------
tools/perf/builtin-kmem.c | 2 +-
tools/perf/builtin-kvm.c | 5 ++--
tools/perf/builtin-kwork.c | 2 +-
tools/perf/builtin-lock.c | 2 +-
tools/perf/builtin-mem.c | 1 -
tools/perf/builtin-record.c | 3 +--
tools/perf/builtin-report.c | 10 +++-----
tools/perf/builtin-sched.c | 4 +--
tools/perf/builtin-script.c | 4 +--
tools/perf/builtin-timechart.c | 2 +-
tools/perf/builtin-trace.c | 2 +-
tools/perf/util/build-id.c | 3 +--
tools/perf/util/build-id.h | 7 +-----
tools/perf/util/data-convert-bt.c | 2 +-
tools/perf/util/data-convert-json.c | 5 ++--
tools/perf/util/intel-tpebs.c | 3 +--
tools/perf/util/jitdump.c | 2 +-
tools/perf/util/session.c | 38 ++++++++++++++---------------
tools/perf/util/tool.c | 4 +--
tools/perf/util/tool.h | 4 +--
24 files changed, 55 insertions(+), 76 deletions(-)
diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index 5e57b78548f4..58e56f826367 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -283,7 +283,6 @@ static int evsel__add_sample(struct evsel *evsel, struct perf_sample *sample,
static int process_sample_event(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct perf_annotate *ann = container_of(tool, struct perf_annotate, tool);
@@ -302,7 +301,7 @@ static int process_sample_event(const struct perf_tool *tool,
goto out_put;
if (!al.filtered &&
- evsel__add_sample(evsel, sample, &al, ann, machine)) {
+ evsel__add_sample(sample->evsel, sample, &al, ann, machine)) {
pr_warning("problem incrementing symbol count, "
"skipping event\n");
ret = -1;
diff --git a/tools/perf/builtin-c2c.c b/tools/perf/builtin-c2c.c
index 72a7802775ee..7593e5908fcc 100644
--- a/tools/perf/builtin-c2c.c
+++ b/tools/perf/builtin-c2c.c
@@ -314,9 +314,9 @@ static void perf_c2c__evsel_hists_inc_stats(struct evsel *evsel,
static int process_sample_event(const struct perf_tool *tool __maybe_unused,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
+ struct evsel *evsel = sample->evsel;
struct c2c_hists *c2c_hists = &c2c.hists;
struct c2c_hist_entry *c2c_he;
struct c2c_stats stats = { .nr_entries = 0, };
diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c
index 1b3df868849a..82e9e514922b 100644
--- a/tools/perf/builtin-diff.c
+++ b/tools/perf/builtin-diff.c
@@ -390,11 +390,11 @@ static struct hist_entry_ops block_hist_ops = {
static int diff__process_sample_event(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct perf_diff *pdiff = container_of(tool, struct perf_diff, tool);
struct addr_location al;
+ struct evsel *evsel = sample->evsel;
struct hists *hists = evsel__hists(evsel);
struct hist_entry_iter iter = {
.evsel = evsel,
diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
index a2493f1097df..f5d1ba497f3c 100644
--- a/tools/perf/builtin-inject.c
+++ b/tools/perf/builtin-inject.c
@@ -363,18 +363,17 @@ typedef int (*inject_handler)(const struct perf_tool *tool,
static int perf_event__repipe_sample(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
- struct perf_inject *inject = container_of(tool, struct perf_inject,
- tool);
+ struct perf_inject *inject = container_of(tool, struct perf_inject, tool);
+ struct evsel *evsel = sample->evsel;
if (evsel && evsel->handler) {
inject_handler f = evsel->handler;
return f(tool, event, sample, evsel, machine);
}
- build_id__mark_dso_hit(tool, event, sample, evsel, machine);
+ build_id__mark_dso_hit(tool, event, sample, machine);
if (inject->itrace_synth_opts.set && sample->aux_sample.size) {
event = perf_inject__cut_auxtrace_sample(inject, event, sample);
@@ -388,10 +387,10 @@ static int perf_event__repipe_sample(const struct perf_tool *tool,
static int perf_event__convert_sample_callchain(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct perf_inject *inject = container_of(tool, struct perf_inject, tool);
+ struct evsel *evsel = sample->evsel;
struct callchain_cursor *cursor = get_tls_callchain_cursor();
union perf_event *event_copy = (void *)inject->event_copy;
struct callchain_cursor_node *node;
@@ -989,10 +988,8 @@ static int mark_dso_hit_callback(struct callchain_cursor_node *node, void *data)
args->mmap_evsel, map, /*sample_in_dso=*/false);
}
-int perf_event__inject_buildid(const struct perf_tool *tool, union perf_event *event,
- struct perf_sample *sample,
- struct evsel *evsel __maybe_unused,
- struct machine *machine)
+static int perf_event__inject_buildid(const struct perf_tool *tool, union perf_event *event,
+ struct perf_sample *sample, struct machine *machine)
{
struct addr_location al;
struct thread *thread;
@@ -1022,7 +1019,7 @@ int perf_event__inject_buildid(const struct perf_tool *tool, union perf_event *e
/*sample_in_dso=*/true);
}
- sample__for_each_callchain_node(thread, evsel, sample, PERF_MAX_STACK_DEPTH,
+ sample__for_each_callchain_node(thread, sample->evsel, sample, PERF_MAX_STACK_DEPTH,
/*symbols=*/false, mark_dso_hit_callback, &args);
thread__put(thread);
repipe:
@@ -1103,7 +1100,7 @@ static int perf_inject__sched_stat(const struct perf_tool *tool,
sample_sw.time = sample->time;
perf_event__synthesize_sample(event_sw, evsel->core.attr.sample_type,
evsel->core.attr.read_format, &sample_sw);
- build_id__mark_dso_hit(tool, event_sw, &sample_sw, evsel, machine);
+ build_id__mark_dso_hit(tool, event_sw, &sample_sw, machine);
ret = perf_event__repipe(tool, event_sw, &sample_sw, machine);
perf_sample__exit(&sample_sw);
return ret;
diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 9c64a0d74823..45b383fa9ce9 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -960,9 +960,9 @@ typedef int (*tracepoint_handler)(struct evsel *evsel,
static int process_sample_event(const struct perf_tool *tool __maybe_unused,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
+ struct evsel *evsel = sample->evsel;
int err = 0;
struct thread *thread = machine__findnew_thread(machine, sample->pid,
sample->tid);
diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c
index 0c5e6b3aac74..5e8e6fde097a 100644
--- a/tools/perf/builtin-kvm.c
+++ b/tools/perf/builtin-kvm.c
@@ -941,9 +941,9 @@ struct vcpu_event_record *per_vcpu_record(struct thread *thread,
static bool handle_kvm_event(struct perf_kvm_stat *kvm,
struct thread *thread,
- struct evsel *evsel,
struct perf_sample *sample)
{
+ struct evsel *evsel = sample->evsel;
struct vcpu_event_record *vcpu_record;
struct event_key key = { .key = INVALID_KEY,
.exit_reasons = kvm->exit_reasons };
@@ -1133,7 +1133,6 @@ static bool skip_sample(struct perf_kvm_stat *kvm,
static int process_sample_event(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
int err = 0;
@@ -1156,7 +1155,7 @@ static int process_sample_event(const struct perf_tool *tool,
return -1;
}
- if (!handle_kvm_event(kvm, thread, evsel, sample))
+ if (!handle_kvm_event(kvm, thread, sample))
err = -1;
thread__put(thread);
diff --git a/tools/perf/builtin-kwork.c b/tools/perf/builtin-kwork.c
index 9d3a4c779a41..d4bb19ed91b5 100644
--- a/tools/perf/builtin-kwork.c
+++ b/tools/perf/builtin-kwork.c
@@ -1955,9 +1955,9 @@ typedef int (*tracepoint_handler)(const struct perf_tool *tool,
static int perf_kwork__process_tracepoint_sample(const struct perf_tool *tool,
union perf_event *event __maybe_unused,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
+ struct evsel *evsel = sample->evsel;
int err = 0;
if (evsel->handler != NULL) {
diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c
index 5585aeb97684..4d8ddf6391e8 100644
--- a/tools/perf/builtin-lock.c
+++ b/tools/perf/builtin-lock.c
@@ -1430,9 +1430,9 @@ typedef int (*tracepoint_handler)(struct evsel *evsel,
static int process_sample_event(const struct perf_tool *tool __maybe_unused,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
+ struct evsel *evsel = sample->evsel;
int err = 0;
struct thread *thread = machine__findnew_thread(machine, sample->pid,
sample->tid);
diff --git a/tools/perf/builtin-mem.c b/tools/perf/builtin-mem.c
index d43500b92a7b..6101a26b3a78 100644
--- a/tools/perf/builtin-mem.c
+++ b/tools/perf/builtin-mem.c
@@ -255,7 +255,6 @@ dump_raw_samples(const struct perf_tool *tool,
static int process_sample_event(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel __maybe_unused,
struct machine *machine)
{
return dump_raw_samples(tool, event, sample, machine);
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 272bba7f4b9e..cc601796b2c8 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -1490,7 +1490,6 @@ static void set_timestamp_boundary(struct record *rec, u64 sample_time)
static int process_sample_event(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct record *rec = container_of(tool, struct record, tool);
@@ -1501,7 +1500,7 @@ static int process_sample_event(const struct perf_tool *tool,
return 0;
rec->samples++;
- return build_id__mark_dso_hit(tool, event, sample, evsel, machine);
+ return build_id__mark_dso_hit(tool, event, sample, machine);
}
static int process_buildids(struct record *rec)
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index 0b0966d94128..9955ce8cce00 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -265,10 +265,10 @@ static int process_feature_event(const struct perf_tool *tool,
static int process_sample_event(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct report *rep = container_of(tool, struct report, tool);
+ struct evsel *evsel = sample->evsel;
struct addr_location al;
struct hist_entry_iter iter = {
.evsel = evsel,
@@ -345,7 +345,6 @@ static int process_sample_event(const struct perf_tool *tool,
static int process_read_event(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample __maybe_unused,
- struct evsel *evsel,
struct machine *machine __maybe_unused)
{
struct report *rep = container_of(tool, struct report, tool);
@@ -353,7 +352,7 @@ static int process_read_event(const struct perf_tool *tool,
if (rep->show_threads) {
int err = perf_read_values_add_value(&rep->show_threads_values,
event->read.pid, event->read.tid,
- evsel,
+ sample->evsel,
event->read.value);
if (err)
@@ -779,11 +778,10 @@ static void report__output_resort(struct report *rep)
static int count_sample_event(const struct perf_tool *tool __maybe_unused,
union perf_event *event __maybe_unused,
- struct perf_sample *sample __maybe_unused,
- struct evsel *evsel,
+ struct perf_sample *sample,
struct machine *machine __maybe_unused)
{
- struct hists *hists = evsel__hists(evsel);
+ struct hists *hists = evsel__hists(sample->evsel);
hists__inc_nr_events(hists);
return 0;
diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
index 53a93aa18853..2950c5c3026e 100644
--- a/tools/perf/builtin-sched.c
+++ b/tools/perf/builtin-sched.c
@@ -1867,9 +1867,9 @@ typedef int (*tracepoint_handler)(const struct perf_tool *tool,
static int perf_sched__process_tracepoint_sample(const struct perf_tool *tool __maybe_unused,
union perf_event *event __maybe_unused,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
+ struct evsel *evsel = sample->evsel;
int err = 0;
if (evsel->handler != NULL) {
@@ -3184,10 +3184,10 @@ typedef int (*sched_handler)(const struct perf_tool *tool,
static int perf_timehist__process_sample(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
+ struct evsel *evsel = sample->evsel;
int err = 0;
struct perf_cpu this_cpu = {
.cpu = sample->cpu,
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index fd0b4609516b..d4e8f49a8751 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -2646,10 +2646,10 @@ static bool filter_cpu(struct perf_sample *sample)
static int process_sample_event(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct perf_script *scr = container_of(tool, struct perf_script, tool);
+ struct evsel *evsel = sample->evsel;
struct addr_location al;
struct addr_location addr_al;
int ret = 0;
@@ -2730,10 +2730,10 @@ static int process_sample_event(const struct perf_tool *tool,
static int process_deferred_sample_event(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct perf_script *scr = container_of(tool, struct perf_script, tool);
+ struct evsel *evsel = sample->evsel;
struct perf_event_attr *attr = &evsel->core.attr;
struct evsel_script *es = evsel->priv;
unsigned int type = output_type(attr->type);
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c
index 28f33e39895d..8692d11ccd29 100644
--- a/tools/perf/builtin-timechart.c
+++ b/tools/perf/builtin-timechart.c
@@ -574,10 +574,10 @@ typedef int (*tracepoint_handler)(struct timechart *tchart,
static int process_sample_event(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct timechart *tchart = container_of(tool, struct timechart, tool);
+ struct evsel *evsel = sample->evsel;
if (evsel->core.attr.sample_type & PERF_SAMPLE_TIME) {
if (!tchart->first_time || tchart->first_time > sample->time)
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index e9cc870a5acd..1aafa77f72cf 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -3501,10 +3501,10 @@ static void trace__set_base_time(struct trace *trace,
static int trace__process_sample(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine __maybe_unused)
{
struct trace *trace = container_of(tool, struct trace, tool);
+ struct evsel *evsel = sample->evsel;
struct thread *thread;
int err = 0;
diff --git a/tools/perf/util/build-id.c b/tools/perf/util/build-id.c
index fdb35133fde4..55b72235f891 100644
--- a/tools/perf/util/build-id.c
+++ b/tools/perf/util/build-id.c
@@ -55,7 +55,6 @@ static int mark_dso_hit_callback(struct callchain_cursor_node *node, void *data
int build_id__mark_dso_hit(const struct perf_tool *tool __maybe_unused,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct addr_location al;
@@ -74,7 +73,7 @@ int build_id__mark_dso_hit(const struct perf_tool *tool __maybe_unused,
addr_location__exit(&al);
- sample__for_each_callchain_node(thread, evsel, sample, PERF_MAX_STACK_DEPTH,
+ sample__for_each_callchain_node(thread, sample->evsel, sample, PERF_MAX_STACK_DEPTH,
/*symbols=*/false, mark_dso_hit_callback, /*data=*/NULL);
diff --git a/tools/perf/util/build-id.h b/tools/perf/util/build-id.h
index 47e621cebe1b..73bad90b06f9 100644
--- a/tools/perf/util/build-id.h
+++ b/tools/perf/util/build-id.h
@@ -34,12 +34,7 @@ char *__dso__build_id_filename(const struct dso *dso, char *bf, size_t size,
bool is_debug, bool is_kallsyms);
int build_id__mark_dso_hit(const struct perf_tool *tool, union perf_event *event,
- struct perf_sample *sample, struct evsel *evsel,
- struct machine *machine);
-
-int perf_event__inject_buildid(const struct perf_tool *tool, union perf_event *event,
- struct perf_sample *sample, struct evsel *evsel,
- struct machine *machine);
+ struct perf_sample *sample, struct machine *machine);
bool perf_session__read_build_ids(struct perf_session *session, bool with_hits);
int perf_session__write_buildid_table(struct perf_session *session,
diff --git a/tools/perf/util/data-convert-bt.c b/tools/perf/util/data-convert-bt.c
index 3b8f2df823a9..b3f745cff2a7 100644
--- a/tools/perf/util/data-convert-bt.c
+++ b/tools/perf/util/data-convert-bt.c
@@ -803,10 +803,10 @@ static bool is_flush_needed(struct ctf_stream *cs)
static int process_sample_event(const struct perf_tool *tool,
union perf_event *_event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine __maybe_unused)
{
struct convert *c = container_of(tool, struct convert, tool);
+ struct evsel *evsel = sample->evsel;
struct evsel_priv *priv = evsel->priv;
struct ctf_writer *cw = &c->writer;
struct ctf_stream *cs;
diff --git a/tools/perf/util/data-convert-json.c b/tools/perf/util/data-convert-json.c
index d526c91312ed..a7da93a7ff0e 100644
--- a/tools/perf/util/data-convert-json.c
+++ b/tools/perf/util/data-convert-json.c
@@ -159,13 +159,12 @@ static void output_sample_callchain_entry(const struct perf_tool *tool,
static int process_sample_event(const struct perf_tool *tool,
union perf_event *event __maybe_unused,
struct perf_sample *sample,
- struct evsel *evsel __maybe_unused,
struct machine *machine)
{
struct convert_json *c = container_of(tool, struct convert_json, tool);
FILE *out = c->out;
struct addr_location al;
- u64 sample_type = __evlist__combined_sample_type(evsel->evlist);
+ u64 sample_type = __evlist__combined_sample_type(sample->evsel->evlist);
u8 cpumode = PERF_RECORD_MISC_USER;
addr_location__init(&al);
@@ -245,7 +244,7 @@ static int process_sample_event(const struct perf_tool *tool,
#ifdef HAVE_LIBTRACEEVENT
if (sample->raw_data) {
- struct tep_event *tp_format = evsel__tp_format(evsel);
+ struct tep_event *tp_format = evsel__tp_format(sample->evsel);
struct tep_format_field **fields = tp_format ? tep_event_fields(tp_format) : NULL;
if (fields) {
diff --git a/tools/perf/util/intel-tpebs.c b/tools/perf/util/intel-tpebs.c
index 8b615dc94e9e..ed8cfe2ba2fa 100644
--- a/tools/perf/util/intel-tpebs.c
+++ b/tools/perf/util/intel-tpebs.c
@@ -185,7 +185,6 @@ static bool should_ignore_sample(const struct perf_sample *sample, const struct
static int process_sample_event(const struct perf_tool *tool __maybe_unused,
union perf_event *event __maybe_unused,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine __maybe_unused)
{
struct tpebs_retire_lat *t;
@@ -196,7 +195,7 @@ static int process_sample_event(const struct perf_tool *tool __maybe_unused,
mutex_unlock(tpebs_mtx_get());
return 0;
}
- t = tpebs_retire_lat__find(evsel);
+ t = tpebs_retire_lat__find(sample->evsel);
if (!t) {
mutex_unlock(tpebs_mtx_get());
return -EINVAL;
diff --git a/tools/perf/util/jitdump.c b/tools/perf/util/jitdump.c
index e0ce8b904729..52e6ffac2b3e 100644
--- a/tools/perf/util/jitdump.c
+++ b/tools/perf/util/jitdump.c
@@ -642,7 +642,7 @@ static int jit_repipe_code_move(struct jit_buf_desc *jd, union jr_entry *jr)
ret = jit_inject_event(jd, event);
if (!ret)
- build_id__mark_dso_hit(tool, event, &sample, NULL, jd->machine);
+ build_id__mark_dso_hit(tool, event, &sample, jd->machine);
out:
perf_sample__exit(&sample);
return ret;
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index fe0de2a0277f..1e25892963b7 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -1117,9 +1117,10 @@ char *get_page_size_name(u64 size, char *str)
return str;
}
-static void dump_sample(struct machine *machine, struct evsel *evsel, union perf_event *event,
+static void dump_sample(struct machine *machine, union perf_event *event,
struct perf_sample *sample)
{
+ struct evsel *evsel = sample->evsel;
u64 sample_type;
char str[PAGE_SIZE_NAME_LEN];
uint16_t e_machine = EM_NONE;
@@ -1183,9 +1184,10 @@ static void dump_sample(struct machine *machine, struct evsel *evsel, union perf
sample_read__printf(sample, evsel->core.attr.read_format);
}
-static void dump_deferred_callchain(struct evsel *evsel, union perf_event *event,
- struct perf_sample *sample)
+static void dump_deferred_callchain(union perf_event *event, struct perf_sample *sample)
{
+ struct evsel *evsel = sample->evsel;
+
if (!dump_trace)
return;
@@ -1291,7 +1293,7 @@ static int deliver_sample_value(struct evlist *evlist,
return 0;
sample->evsel = container_of(sid->evsel, struct evsel, core);
- ret = tool->sample(tool, event, sample, sample->evsel, machine);
+ ret = tool->sample(tool, event, sample, machine);
sample->evsel = saved_evsel;
return ret;
}
@@ -1323,8 +1325,9 @@ static int deliver_sample_group(struct evlist *evlist,
static int evlist__deliver_sample(struct evlist *evlist, const struct perf_tool *tool,
union perf_event *event, struct perf_sample *sample,
- struct evsel *evsel, struct machine *machine)
+ struct machine *machine)
{
+ struct evsel *evsel = sample->evsel;
/* We know evsel != NULL. */
u64 sample_type = evsel->core.attr.sample_type;
u64 read_format = evsel->core.attr.read_format;
@@ -1332,7 +1335,7 @@ static int evlist__deliver_sample(struct evlist *evlist, const struct perf_tool
/* Standard sample delivery. */
if (!(sample_type & PERF_SAMPLE_READ))
- return tool->sample(tool, event, sample, evsel, machine);
+ return tool->sample(tool, event, sample, machine);
/* For PERF_SAMPLE_READ we have either single or group mode. */
if (read_format & PERF_FORMAT_GROUP)
@@ -1371,8 +1374,7 @@ static int evlist__deliver_deferred_callchain(struct evlist *evlist,
struct evsel *saved_evsel = sample->evsel;
sample->evsel = evlist__id2evsel(evlist, sample->id);
- ret = tool->callchain_deferred(tool, event, sample,
- sample->evsel, machine);
+ ret = tool->callchain_deferred(tool, event, sample, machine);
sample->evsel = saved_evsel;
return ret;
}
@@ -1400,7 +1402,7 @@ static int evlist__deliver_deferred_callchain(struct evlist *evlist,
orig_sample.evsel = evlist__id2evsel(evlist, orig_sample.id);
ret = evlist__deliver_sample(evlist, tool, de->event,
- &orig_sample, orig_sample.evsel, machine);
+ &orig_sample, machine);
perf_sample__exit(&orig_sample);
list_del(&de->list);
@@ -1438,7 +1440,7 @@ static int session__flush_deferred_samples(struct perf_session *session,
sample.evsel = evlist__id2evsel(evlist, sample.id);
ret = evlist__deliver_sample(evlist, tool, de->event,
- &sample, sample.evsel, machine);
+ &sample, machine);
perf_sample__exit(&sample);
list_del(&de->list);
@@ -1458,7 +1460,6 @@ static int machines__deliver_event(struct machines *machines,
const struct perf_tool *tool, u64 file_offset,
const char *file_path)
{
- struct evsel *evsel;
struct machine *machine;
dump_event(evlist, event, file_offset, sample, file_path);
@@ -1468,21 +1469,20 @@ static int machines__deliver_event(struct machines *machines,
else
assert(sample->evsel == evlist__id2evsel(evlist, sample->id));
- evsel = sample->evsel;
machine = machines__find_for_cpumode(machines, event, sample);
switch (event->header.type) {
case PERF_RECORD_SAMPLE:
- if (evsel == NULL) {
+ if (sample->evsel == NULL) {
++evlist->stats.nr_unknown_id;
return 0;
}
if (machine == NULL) {
++evlist->stats.nr_unprocessable_samples;
- dump_sample(machine, evsel, event, sample);
+ dump_sample(machine, event, sample);
return 0;
}
- dump_sample(machine, evsel, event, sample);
+ dump_sample(machine, event, sample);
if (sample->deferred_callchain && tool->merge_deferred_callchains) {
struct deferred_event *de = malloc(sizeof(*de));
size_t sz = event->header.size;
@@ -1499,7 +1499,7 @@ static int machines__deliver_event(struct machines *machines,
list_add_tail(&de->list, &evlist->deferred_samples);
return 0;
}
- return evlist__deliver_sample(evlist, tool, event, sample, evsel, machine);
+ return evlist__deliver_sample(evlist, tool, event, sample, machine);
case PERF_RECORD_MMAP:
return tool->mmap(tool, event, sample, machine);
case PERF_RECORD_MMAP2:
@@ -1527,8 +1527,8 @@ static int machines__deliver_event(struct machines *machines,
evlist->stats.total_lost_samples += event->lost_samples.lost;
return tool->lost_samples(tool, event, sample, machine);
case PERF_RECORD_READ:
- dump_read(evsel, event);
- return tool->read(tool, event, sample, evsel, machine);
+ dump_read(sample->evsel, event);
+ return tool->read(tool, event, sample, machine);
case PERF_RECORD_THROTTLE:
return tool->throttle(tool, event, sample, machine);
case PERF_RECORD_UNTHROTTLE:
@@ -1557,7 +1557,7 @@ static int machines__deliver_event(struct machines *machines,
case PERF_RECORD_AUX_OUTPUT_HW_ID:
return tool->aux_output_hw_id(tool, event, sample, machine);
case PERF_RECORD_CALLCHAIN_DEFERRED:
- dump_deferred_callchain(evsel, event, sample);
+ dump_deferred_callchain(event, sample);
return evlist__deliver_deferred_callchain(evlist, tool, event,
sample, machine);
default:
diff --git a/tools/perf/util/tool.c b/tools/perf/util/tool.c
index ff2150517b75..225a77d530ce 100644
--- a/tools/perf/util/tool.c
+++ b/tools/perf/util/tool.c
@@ -110,7 +110,6 @@ static int process_event_synth_event_update_stub(const struct perf_tool *tool __
int process_event_sample_stub(const struct perf_tool *tool __maybe_unused,
union perf_event *event __maybe_unused,
struct perf_sample *sample __maybe_unused,
- struct evsel *evsel __maybe_unused,
struct machine *machine __maybe_unused)
{
dump_printf(": unhandled!\n");
@@ -349,12 +348,11 @@ bool perf_tool__compressed_is_stub(const struct perf_tool *tool)
static int delegate_ ## name(const struct perf_tool *tool, \
union perf_event *event, \
struct perf_sample *sample, \
- struct evsel *evsel, \
struct machine *machine) \
{ \
struct delegate_tool *del_tool = container_of(tool, struct delegate_tool, tool); \
struct perf_tool *delegate = del_tool->delegate; \
- return delegate->name(delegate, event, sample, evsel, machine); \
+ return delegate->name(delegate, event, sample, machine); \
}
CREATE_DELEGATE_SAMPLE(read);
CREATE_DELEGATE_SAMPLE(sample);
diff --git a/tools/perf/util/tool.h b/tools/perf/util/tool.h
index 2d9a4b1ca9d0..2a4f124ffd8d 100644
--- a/tools/perf/util/tool.h
+++ b/tools/perf/util/tool.h
@@ -9,7 +9,6 @@
struct perf_session;
union perf_event;
struct evlist;
-struct evsel;
struct perf_sample;
struct perf_tool;
struct machine;
@@ -17,7 +16,7 @@ struct ordered_events;
typedef int (*event_sample)(const struct perf_tool *tool, union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel, struct machine *machine);
+ struct machine *machine);
typedef int (*event_op)(const struct perf_tool *tool, union perf_event *event,
struct perf_sample *sample, struct machine *machine);
@@ -103,7 +102,6 @@ bool perf_tool__compressed_is_stub(const struct perf_tool *tool);
int process_event_sample_stub(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine);
struct delegate_tool {
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from kvm-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-kvm.c | 19 +++--
.../perf/util/kvm-stat-arch/kvm-stat-arm64.c | 17 +++--
.../util/kvm-stat-arch/kvm-stat-loongarch.c | 17 ++---
.../util/kvm-stat-arch/kvm-stat-powerpc.c | 17 ++---
.../perf/util/kvm-stat-arch/kvm-stat-riscv.c | 18 +++--
tools/perf/util/kvm-stat-arch/kvm-stat-s390.c | 20 +++---
tools/perf/util/kvm-stat-arch/kvm-stat-x86.c | 70 ++++++++-----------
tools/perf/util/kvm-stat.c | 19 +++--
tools/perf/util/kvm-stat.h | 18 ++---
9 files changed, 89 insertions(+), 126 deletions(-)
diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c
index 5e8e6fde097a..d9b9792894a8 100644
--- a/tools/perf/builtin-kvm.c
+++ b/tools/perf/builtin-kvm.c
@@ -806,7 +806,6 @@ static bool update_kvm_event(struct perf_kvm_stat *kvm,
}
static bool is_child_event(struct perf_kvm_stat *kvm,
- struct evsel *evsel,
struct perf_sample *sample,
struct event_key *key)
{
@@ -818,8 +817,8 @@ static bool is_child_event(struct perf_kvm_stat *kvm,
return false;
for (; child_ops->name; child_ops++) {
- if (evsel__name_is(evsel, child_ops->name)) {
- child_ops->get_key(evsel, sample, key);
+ if (evsel__name_is(sample->evsel, child_ops->name)) {
+ child_ops->get_key(sample, key);
return true;
}
}
@@ -917,11 +916,10 @@ static bool handle_end_event(struct perf_kvm_stat *kvm,
static
struct vcpu_event_record *per_vcpu_record(struct thread *thread,
- struct evsel *evsel,
struct perf_sample *sample)
{
/* Only kvm_entry records vcpu id. */
- if (!thread__priv(thread) && kvm_entry_event(evsel)) {
+ if (!thread__priv(thread) && kvm_entry_event(sample->evsel)) {
struct vcpu_event_record *vcpu_record;
struct machine *machine = maps__machine(thread__maps(thread));
uint16_t e_machine = thread__e_machine(thread, machine, /*e_flags=*/NULL);
@@ -932,7 +930,7 @@ struct vcpu_event_record *per_vcpu_record(struct thread *thread,
return NULL;
}
- vcpu_record->vcpu_id = evsel__intval(evsel, sample, vcpu_id_str(e_machine));
+ vcpu_record->vcpu_id = evsel__intval(sample->evsel, sample, vcpu_id_str(e_machine));
thread__set_priv(thread, vcpu_record);
}
@@ -943,12 +941,11 @@ static bool handle_kvm_event(struct perf_kvm_stat *kvm,
struct thread *thread,
struct perf_sample *sample)
{
- struct evsel *evsel = sample->evsel;
struct vcpu_event_record *vcpu_record;
struct event_key key = { .key = INVALID_KEY,
.exit_reasons = kvm->exit_reasons };
- vcpu_record = per_vcpu_record(thread, evsel, sample);
+ vcpu_record = per_vcpu_record(thread, sample);
if (!vcpu_record)
return true;
@@ -957,13 +954,13 @@ static bool handle_kvm_event(struct perf_kvm_stat *kvm,
(kvm->trace_vcpu != vcpu_record->vcpu_id))
return true;
- if (kvm->events_ops->is_begin_event(evsel, sample, &key))
+ if (kvm->events_ops->is_begin_event(sample, &key))
return handle_begin_event(kvm, vcpu_record, &key, sample);
- if (is_child_event(kvm, evsel, sample, &key))
+ if (is_child_event(kvm, sample, &key))
return handle_child_event(kvm, vcpu_record, &key, sample);
- if (kvm->events_ops->is_end_event(evsel, sample, &key))
+ if (kvm->events_ops->is_end_event(sample, &key))
return handle_end_event(kvm, vcpu_record, &key, sample);
return true;
diff --git a/tools/perf/util/kvm-stat-arch/kvm-stat-arm64.c b/tools/perf/util/kvm-stat-arch/kvm-stat-arm64.c
index c640dcd8af7c..1e76906f719c 100644
--- a/tools/perf/util/kvm-stat-arch/kvm-stat-arm64.c
+++ b/tools/perf/util/kvm-stat-arch/kvm-stat-arm64.c
@@ -17,10 +17,11 @@ static const char * const __kvm_events_tp[] = {
NULL,
};
-static void event_get_key(struct evsel *evsel,
- struct perf_sample *sample,
+static void event_get_key(struct perf_sample *sample,
struct event_key *key)
{
+ struct evsel *evsel = sample->evsel;
+
key->info = 0;
key->key = evsel__intval(evsel, sample, kvm_exit_reason(EM_AARCH64));
key->exit_reasons = arm64_exit_reasons;
@@ -36,19 +37,17 @@ static void event_get_key(struct evsel *evsel,
}
}
-static bool event_begin(struct evsel *evsel,
- struct perf_sample *sample __maybe_unused,
+static bool event_begin(struct perf_sample *sample,
struct event_key *key __maybe_unused)
{
- return evsel__name_is(evsel, kvm_entry_trace(EM_AARCH64));
+ return evsel__name_is(sample->evsel, kvm_entry_trace(EM_AARCH64));
}
-static bool event_end(struct evsel *evsel,
- struct perf_sample *sample,
+static bool event_end(struct perf_sample *sample,
struct event_key *key)
{
- if (evsel__name_is(evsel, kvm_exit_trace(EM_AARCH64))) {
- event_get_key(evsel, sample, key);
+ if (evsel__name_is(sample->evsel, kvm_exit_trace(EM_AARCH64))) {
+ event_get_key(sample, key);
return true;
}
return false;
diff --git a/tools/perf/util/kvm-stat-arch/kvm-stat-loongarch.c b/tools/perf/util/kvm-stat-arch/kvm-stat-loongarch.c
index b802e516b138..9d6265290f6d 100644
--- a/tools/perf/util/kvm-stat-arch/kvm-stat-loongarch.c
+++ b/tools/perf/util/kvm-stat-arch/kvm-stat-loongarch.c
@@ -53,14 +53,12 @@ static const char * const __kvm_events_tp[] = {
NULL,
};
-static bool event_begin(struct evsel *evsel,
- struct perf_sample *sample, struct event_key *key)
+static bool event_begin(struct perf_sample *sample, struct event_key *key)
{
- return exit_event_begin(evsel, sample, key);
+ return exit_event_begin(sample, key);
}
-static bool event_end(struct evsel *evsel,
- struct perf_sample *sample __maybe_unused,
+static bool event_end(struct perf_sample *sample,
struct event_key *key __maybe_unused)
{
/*
@@ -71,17 +69,16 @@ static bool event_end(struct evsel *evsel,
* kvm:kvm_enter means returning to vmm and then to guest
* kvm:kvm_reenter means returning to guest immediately
*/
- return evsel__name_is(evsel, kvm_entry_trace(EM_LOONGARCH)) ||
- evsel__name_is(evsel, kvm_reenter_trace);
+ return evsel__name_is(sample->evsel, kvm_entry_trace(EM_LOONGARCH)) ||
+ evsel__name_is(sample->evsel, kvm_reenter_trace);
}
-static void event_gspr_get_key(struct evsel *evsel,
- struct perf_sample *sample, struct event_key *key)
+static void event_gspr_get_key(struct perf_sample *sample, struct event_key *key)
{
unsigned int insn;
key->key = LOONGARCH_EXCEPTION_OTHERS;
- insn = evsel__intval(evsel, sample, "inst_word");
+ insn = evsel__intval(sample->evsel, sample, "inst_word");
switch (insn >> 24) {
case 0:
diff --git a/tools/perf/util/kvm-stat-arch/kvm-stat-powerpc.c b/tools/perf/util/kvm-stat-arch/kvm-stat-powerpc.c
index 42182d70beb6..5158d7e88ee6 100644
--- a/tools/perf/util/kvm-stat-arch/kvm-stat-powerpc.c
+++ b/tools/perf/util/kvm-stat-arch/kvm-stat-powerpc.c
@@ -28,12 +28,11 @@ static const char * const ppc_book3s_hv_kvm_tp[] = {
/* 1 extra placeholder for NULL */
static const char *__kvm_events_tp[NR_TPS + 1];
-static void hcall_event_get_key(struct evsel *evsel,
- struct perf_sample *sample,
+static void hcall_event_get_key(struct perf_sample *sample,
struct event_key *key)
{
key->info = 0;
- key->key = evsel__intval(evsel, sample, "req");
+ key->key = evsel__intval(sample->evsel, sample, "req");
}
static const char *get_hcall_exit_reason(u64 exit_code)
@@ -51,18 +50,16 @@ static const char *get_hcall_exit_reason(u64 exit_code)
return "UNKNOWN";
}
-static bool hcall_event_end(struct evsel *evsel,
- struct perf_sample *sample __maybe_unused,
+static bool hcall_event_end(struct perf_sample *sample,
struct event_key *key __maybe_unused)
{
- return evsel__name_is(evsel, __kvm_events_tp[3]);
+ return evsel__name_is(sample->evsel, __kvm_events_tp[3]);
}
-static bool hcall_event_begin(struct evsel *evsel,
- struct perf_sample *sample, struct event_key *key)
+static bool hcall_event_begin(struct perf_sample *sample, struct event_key *key)
{
- if (evsel__name_is(evsel, __kvm_events_tp[2])) {
- hcall_event_get_key(evsel, sample, key);
+ if (evsel__name_is(sample->evsel, __kvm_events_tp[2])) {
+ hcall_event_get_key(sample, key);
return true;
}
diff --git a/tools/perf/util/kvm-stat-arch/kvm-stat-riscv.c b/tools/perf/util/kvm-stat-arch/kvm-stat-riscv.c
index 8d4d5d6ce720..e8db8b4f8e2e 100644
--- a/tools/perf/util/kvm-stat-arch/kvm-stat-riscv.c
+++ b/tools/perf/util/kvm-stat-arch/kvm-stat-riscv.c
@@ -20,30 +20,28 @@ static const char * const __kvm_events_tp[] = {
NULL,
};
-static void event_get_key(struct evsel *evsel,
- struct perf_sample *sample,
+static void event_get_key(struct perf_sample *sample,
struct event_key *key)
{
int xlen = 64; // TODO: 32-bit support.
key->info = 0;
- key->key = evsel__intval(evsel, sample, kvm_exit_reason(EM_RISCV)) & ~CAUSE_IRQ_FLAG(xlen);
+ key->key = evsel__intval(sample->evsel, sample,
+ kvm_exit_reason(EM_RISCV)) & ~CAUSE_IRQ_FLAG(xlen);
key->exit_reasons = riscv_exit_reasons;
}
-static bool event_begin(struct evsel *evsel,
- struct perf_sample *sample __maybe_unused,
+static bool event_begin(struct perf_sample *sample,
struct event_key *key __maybe_unused)
{
- return evsel__name_is(evsel, kvm_entry_trace(EM_RISCV));
+ return evsel__name_is(sample->evsel, kvm_entry_trace(EM_RISCV));
}
-static bool event_end(struct evsel *evsel,
- struct perf_sample *sample,
+static bool event_end(struct perf_sample *sample,
struct event_key *key)
{
- if (evsel__name_is(evsel, kvm_exit_trace(EM_RISCV))) {
- event_get_key(evsel, sample, key);
+ if (evsel__name_is(sample->evsel, kvm_exit_trace(EM_RISCV))) {
+ event_get_key(sample, key);
return true;
}
return false;
diff --git a/tools/perf/util/kvm-stat-arch/kvm-stat-s390.c b/tools/perf/util/kvm-stat-arch/kvm-stat-s390.c
index 7e29169f5bb0..158372ba0205 100644
--- a/tools/perf/util/kvm-stat-arch/kvm-stat-s390.c
+++ b/tools/perf/util/kvm-stat-arch/kvm-stat-s390.c
@@ -18,38 +18,34 @@ define_exit_reasons_table(sie_sigp_order_codes, sigp_order_codes);
define_exit_reasons_table(sie_diagnose_codes, diagnose_codes);
define_exit_reasons_table(sie_icpt_prog_codes, icpt_prog_codes);
-static void event_icpt_insn_get_key(struct evsel *evsel,
- struct perf_sample *sample,
+static void event_icpt_insn_get_key(struct perf_sample *sample,
struct event_key *key)
{
u64 insn;
- insn = evsel__intval(evsel, sample, "instruction");
+ insn = evsel__intval(sample->evsel, sample, "instruction");
key->key = icpt_insn_decoder(insn);
key->exit_reasons = sie_icpt_insn_codes;
}
-static void event_sigp_get_key(struct evsel *evsel,
- struct perf_sample *sample,
+static void event_sigp_get_key(struct perf_sample *sample,
struct event_key *key)
{
- key->key = evsel__intval(evsel, sample, "order_code");
+ key->key = evsel__intval(sample->evsel, sample, "order_code");
key->exit_reasons = sie_sigp_order_codes;
}
-static void event_diag_get_key(struct evsel *evsel,
- struct perf_sample *sample,
+static void event_diag_get_key(struct perf_sample *sample,
struct event_key *key)
{
- key->key = evsel__intval(evsel, sample, "code");
+ key->key = evsel__intval(sample->evsel, sample, "code");
key->exit_reasons = sie_diagnose_codes;
}
-static void event_icpt_prog_get_key(struct evsel *evsel,
- struct perf_sample *sample,
+static void event_icpt_prog_get_key(struct perf_sample *sample,
struct event_key *key)
{
- key->key = evsel__intval(evsel, sample, "code");
+ key->key = evsel__intval(sample->evsel, sample, "code");
key->exit_reasons = sie_icpt_prog_codes;
}
diff --git a/tools/perf/util/kvm-stat-arch/kvm-stat-x86.c b/tools/perf/util/kvm-stat-arch/kvm-stat-x86.c
index 0f626db3a439..0ce543d82850 100644
--- a/tools/perf/util/kvm-stat-arch/kvm-stat-x86.c
+++ b/tools/perf/util/kvm-stat-arch/kvm-stat-x86.c
@@ -24,45 +24,43 @@ static const struct kvm_events_ops exit_events = {
* the time of MMIO write: kvm_mmio(KVM_TRACE_MMIO_WRITE...) -> kvm_entry
* the time of MMIO read: kvm_exit -> kvm_mmio(KVM_TRACE_MMIO_READ...).
*/
-static void mmio_event_get_key(struct evsel *evsel, struct perf_sample *sample,
+static void mmio_event_get_key(struct perf_sample *sample,
struct event_key *key)
{
- key->key = evsel__intval(evsel, sample, "gpa");
- key->info = evsel__intval(evsel, sample, "type");
+ key->key = evsel__intval(sample->evsel, sample, "gpa");
+ key->info = evsel__intval(sample->evsel, sample, "type");
}
#define KVM_TRACE_MMIO_READ_UNSATISFIED 0
#define KVM_TRACE_MMIO_READ 1
#define KVM_TRACE_MMIO_WRITE 2
-static bool mmio_event_begin(struct evsel *evsel,
- struct perf_sample *sample, struct event_key *key)
+static bool mmio_event_begin(struct perf_sample *sample, struct event_key *key)
{
/* MMIO read begin event in kernel. */
- if (kvm_exit_event(evsel))
+ if (kvm_exit_event(sample->evsel))
return true;
/* MMIO write begin event in kernel. */
- if (evsel__name_is(evsel, "kvm:kvm_mmio") &&
- evsel__intval(evsel, sample, "type") == KVM_TRACE_MMIO_WRITE) {
- mmio_event_get_key(evsel, sample, key);
+ if (evsel__name_is(sample->evsel, "kvm:kvm_mmio") &&
+ evsel__intval(sample->evsel, sample, "type") == KVM_TRACE_MMIO_WRITE) {
+ mmio_event_get_key(sample, key);
return true;
}
return false;
}
-static bool mmio_event_end(struct evsel *evsel, struct perf_sample *sample,
- struct event_key *key)
+static bool mmio_event_end(struct perf_sample *sample, struct event_key *key)
{
/* MMIO write end event in kernel. */
- if (kvm_entry_event(evsel))
+ if (kvm_entry_event(sample->evsel))
return true;
/* MMIO read end event in kernel.*/
- if (evsel__name_is(evsel, "kvm:kvm_mmio") &&
- evsel__intval(evsel, sample, "type") == KVM_TRACE_MMIO_READ) {
- mmio_event_get_key(evsel, sample, key);
+ if (evsel__name_is(sample->evsel, "kvm:kvm_mmio") &&
+ evsel__intval(sample->evsel, sample, "type") == KVM_TRACE_MMIO_READ) {
+ mmio_event_get_key(sample, key);
return true;
}
@@ -86,31 +84,27 @@ static const struct kvm_events_ops mmio_events = {
};
/* The time of emulation pio access is from kvm_pio to kvm_entry. */
-static void ioport_event_get_key(struct evsel *evsel,
- struct perf_sample *sample,
+static void ioport_event_get_key(struct perf_sample *sample,
struct event_key *key)
{
- key->key = evsel__intval(evsel, sample, "port");
- key->info = evsel__intval(evsel, sample, "rw");
+ key->key = evsel__intval(sample->evsel, sample, "port");
+ key->info = evsel__intval(sample->evsel, sample, "rw");
}
-static bool ioport_event_begin(struct evsel *evsel,
- struct perf_sample *sample,
+static bool ioport_event_begin(struct perf_sample *sample,
struct event_key *key)
{
- if (evsel__name_is(evsel, "kvm:kvm_pio")) {
- ioport_event_get_key(evsel, sample, key);
+ if (evsel__name_is(sample->evsel, "kvm:kvm_pio")) {
+ ioport_event_get_key(sample, key);
return true;
}
return false;
}
-static bool ioport_event_end(struct evsel *evsel,
- struct perf_sample *sample __maybe_unused,
- struct event_key *key __maybe_unused)
+static bool ioport_event_end(struct perf_sample *sample, struct event_key *key __maybe_unused)
{
- return kvm_entry_event(evsel);
+ return kvm_entry_event(sample->evsel);
}
static void ioport_event_decode_key(struct perf_kvm_stat *kvm __maybe_unused,
@@ -130,31 +124,25 @@ static const struct kvm_events_ops ioport_events = {
};
/* The time of emulation msr is from kvm_msr to kvm_entry. */
-static void msr_event_get_key(struct evsel *evsel,
- struct perf_sample *sample,
- struct event_key *key)
+static void msr_event_get_key(struct perf_sample *sample, struct event_key *key)
{
- key->key = evsel__intval(evsel, sample, "ecx");
- key->info = evsel__intval(evsel, sample, "write");
+ key->key = evsel__intval(sample->evsel, sample, "ecx");
+ key->info = evsel__intval(sample->evsel, sample, "write");
}
-static bool msr_event_begin(struct evsel *evsel,
- struct perf_sample *sample,
- struct event_key *key)
+static bool msr_event_begin(struct perf_sample *sample, struct event_key *key)
{
- if (evsel__name_is(evsel, "kvm:kvm_msr")) {
- msr_event_get_key(evsel, sample, key);
+ if (evsel__name_is(sample->evsel, "kvm:kvm_msr")) {
+ msr_event_get_key(sample, key);
return true;
}
return false;
}
-static bool msr_event_end(struct evsel *evsel,
- struct perf_sample *sample __maybe_unused,
- struct event_key *key __maybe_unused)
+static bool msr_event_end(struct perf_sample *sample, struct event_key *key __maybe_unused)
{
- return kvm_entry_event(evsel);
+ return kvm_entry_event(sample->evsel);
}
static void msr_event_decode_key(struct perf_kvm_stat *kvm __maybe_unused,
diff --git a/tools/perf/util/kvm-stat.c b/tools/perf/util/kvm-stat.c
index 27f16810498c..f17a6132958d 100644
--- a/tools/perf/util/kvm-stat.c
+++ b/tools/perf/util/kvm-stat.c
@@ -11,22 +11,20 @@ bool kvm_exit_event(struct evsel *evsel)
return evsel__name_is(evsel, kvm_exit_trace(e_machine));
}
-void exit_event_get_key(struct evsel *evsel,
- struct perf_sample *sample,
+void exit_event_get_key(struct perf_sample *sample,
struct event_key *key)
{
- uint16_t e_machine = evsel__e_machine(evsel, /*e_flags=*/NULL);
+ uint16_t e_machine = evsel__e_machine(sample->evsel, /*e_flags=*/NULL);
key->info = 0;
- key->key = evsel__intval(evsel, sample, kvm_exit_reason(e_machine));
+ key->key = evsel__intval(sample->evsel, sample, kvm_exit_reason(e_machine));
}
-bool exit_event_begin(struct evsel *evsel,
- struct perf_sample *sample, struct event_key *key)
+bool exit_event_begin(struct perf_sample *sample, struct event_key *key)
{
- if (kvm_exit_event(evsel)) {
- exit_event_get_key(evsel, sample, key);
+ if (kvm_exit_event(sample->evsel)) {
+ exit_event_get_key(sample, key);
return true;
}
@@ -40,11 +38,10 @@ bool kvm_entry_event(struct evsel *evsel)
return evsel__name_is(evsel, kvm_entry_trace(e_machine));
}
-bool exit_event_end(struct evsel *evsel,
- struct perf_sample *sample __maybe_unused,
+bool exit_event_end(struct perf_sample *sample,
struct event_key *key __maybe_unused)
{
- return kvm_entry_event(evsel);
+ return kvm_entry_event(sample->evsel);
}
static const char *get_exit_reason(struct perf_kvm_stat *kvm,
diff --git a/tools/perf/util/kvm-stat.h b/tools/perf/util/kvm-stat.h
index 4a998aaece5d..cdbd921a555f 100644
--- a/tools/perf/util/kvm-stat.h
+++ b/tools/perf/util/kvm-stat.h
@@ -53,18 +53,15 @@ struct kvm_event {
};
struct child_event_ops {
- void (*get_key)(struct evsel *evsel,
- struct perf_sample *sample,
+ void (*get_key)(struct perf_sample *sample,
struct event_key *key);
const char *name;
};
struct kvm_events_ops {
- bool (*is_begin_event)(struct evsel *evsel,
- struct perf_sample *sample,
+ bool (*is_begin_event)(struct perf_sample *sample,
struct event_key *key);
- bool (*is_end_event)(struct evsel *evsel,
- struct perf_sample *sample, struct event_key *key);
+ bool (*is_end_event)(struct perf_sample *sample, struct event_key *key);
const struct child_event_ops *child_ops;
void (*decode_key)(struct perf_kvm_stat *kvm, struct event_key *key,
char *decode);
@@ -116,14 +113,11 @@ struct kvm_reg_events_ops {
#ifdef HAVE_LIBTRACEEVENT
-void exit_event_get_key(struct evsel *evsel,
- struct perf_sample *sample,
+void exit_event_get_key(struct perf_sample *sample,
struct event_key *key);
-bool exit_event_begin(struct evsel *evsel,
- struct perf_sample *sample,
+bool exit_event_begin(struct perf_sample *sample,
struct event_key *key);
-bool exit_event_end(struct evsel *evsel,
- struct perf_sample *sample,
+bool exit_event_end(struct perf_sample *sample,
struct event_key *key);
void exit_event_decode_key(struct perf_kvm_stat *kvm,
struct event_key *key,
--
2.54.0.746.g67dd491aae-googThe evsel argument to evsel__intval, evsel__rawptr, and similar
functions, is unnecessary as it can be read from the sample. Remove
the evsel and rename the function to match that the data is coming
from the sample.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-inject.c | 2 +-
tools/perf/builtin-kmem.c | 34 +++----
tools/perf/builtin-kvm.c | 2 +-
tools/perf/builtin-kwork.c | 29 +++---
tools/perf/builtin-lock.c | 32 +++----
tools/perf/builtin-sched.c | 92 +++++++++----------
tools/perf/builtin-timechart.c | 88 +++++++++---------
tools/perf/builtin-trace.c | 12 +--
tools/perf/tests/openat-syscall-tp-fields.c | 2 +-
tools/perf/tests/switch-tracking.c | 4 +-
tools/perf/util/evsel.c | 22 ++---
tools/perf/util/evsel.h | 12 +--
tools/perf/util/intel-pt.c | 2 +-
.../perf/util/kvm-stat-arch/kvm-stat-arm64.c | 6 +-
.../util/kvm-stat-arch/kvm-stat-loongarch.c | 2 +-
.../util/kvm-stat-arch/kvm-stat-powerpc.c | 2 +-
.../perf/util/kvm-stat-arch/kvm-stat-riscv.c | 3 +-
tools/perf/util/kvm-stat-arch/kvm-stat-s390.c | 8 +-
tools/perf/util/kvm-stat-arch/kvm-stat-x86.c | 16 ++--
tools/perf/util/kvm-stat.c | 2 +-
20 files changed, 183 insertions(+), 189 deletions(-)
diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
index f5d1ba497f3c..3dac8bcaf6a1 100644
--- a/tools/perf/builtin-inject.c
+++ b/tools/perf/builtin-inject.c
@@ -1083,7 +1083,7 @@ static int perf_inject__sched_stat(const struct perf_tool *tool,
union perf_event *event_sw;
struct perf_sample sample_sw;
struct perf_inject *inject = container_of(tool, struct perf_inject, tool);
- u32 pid = evsel__intval(evsel, sample, "pid");
+ u32 pid = perf_sample__intval(sample, "pid");
int ret;
list_for_each_entry(ent, &inject->samples, node) {
diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 45b383fa9ce9..d7233a8b43be 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -173,10 +173,10 @@ static int insert_caller_stat(unsigned long call_site,
static int evsel__process_alloc_event(struct evsel *evsel, struct perf_sample *sample)
{
- unsigned long ptr = evsel__intval(evsel, sample, "ptr"),
- call_site = evsel__intval(evsel, sample, "call_site");
- int bytes_req = evsel__intval(evsel, sample, "bytes_req"),
- bytes_alloc = evsel__intval(evsel, sample, "bytes_alloc");
+ unsigned long ptr = perf_sample__intval(sample, "ptr"),
+ call_site = perf_sample__intval(sample, "call_site");
+ int bytes_req = perf_sample__intval(sample, "bytes_req"),
+ bytes_alloc = perf_sample__intval(sample, "bytes_alloc");
if (insert_alloc_stat(call_site, ptr, bytes_req, bytes_alloc, sample->cpu) ||
insert_caller_stat(call_site, bytes_req, bytes_alloc))
@@ -202,7 +202,7 @@ static int evsel__process_alloc_event(struct evsel *evsel, struct perf_sample *s
int node1, node2;
node1 = cpu__get_node((struct perf_cpu){.cpu = sample->cpu});
- node2 = evsel__intval(evsel, sample, "node");
+ node2 = perf_sample__intval(sample, "node");
/*
* If the field "node" is NUMA_NO_NODE (-1), we don't take it
@@ -243,9 +243,9 @@ static struct alloc_stat *search_alloc_stat(unsigned long ptr,
return NULL;
}
-static int evsel__process_free_event(struct evsel *evsel, struct perf_sample *sample)
+static int evsel__process_free_event(struct evsel *evsel __maybe_unused, struct perf_sample *sample)
{
- unsigned long ptr = evsel__intval(evsel, sample, "ptr");
+ unsigned long ptr = perf_sample__intval(sample, "ptr");
struct alloc_stat *s_alloc, *s_caller;
s_alloc = search_alloc_stat(ptr, 0, &root_alloc_stat, ptr_cmp);
@@ -808,10 +808,9 @@ static int parse_gfp_flags(struct evsel *evsel, struct perf_sample *sample,
static int evsel__process_page_alloc_event(struct evsel *evsel, struct perf_sample *sample)
{
u64 page;
- unsigned int order = evsel__intval(evsel, sample, "order");
- unsigned int gfp_flags = evsel__intval(evsel, sample, "gfp_flags");
- unsigned int migrate_type = evsel__intval(evsel, sample,
- "migratetype");
+ unsigned int order = perf_sample__intval(sample, "order");
+ unsigned int gfp_flags = perf_sample__intval(sample, "gfp_flags");
+ unsigned int migrate_type = perf_sample__intval(sample, "migratetype");
u64 bytes = kmem_page_size << order;
u64 callsite;
struct page_stat *pstat;
@@ -822,9 +821,9 @@ static int evsel__process_page_alloc_event(struct evsel *evsel, struct perf_samp
};
if (use_pfn)
- page = evsel__intval(evsel, sample, "pfn");
+ page = perf_sample__intval(sample, "pfn");
else
- page = evsel__intval(evsel, sample, "page");
+ page = perf_sample__intval(sample, "page");
nr_page_allocs++;
total_page_alloc_bytes += bytes;
@@ -877,10 +876,11 @@ static int evsel__process_page_alloc_event(struct evsel *evsel, struct perf_samp
return 0;
}
-static int evsel__process_page_free_event(struct evsel *evsel, struct perf_sample *sample)
+static int evsel__process_page_free_event(struct evsel *evsel __maybe_unused,
+ struct perf_sample *sample)
{
u64 page;
- unsigned int order = evsel__intval(evsel, sample, "order");
+ unsigned int order = perf_sample__intval(sample, "order");
u64 bytes = kmem_page_size << order;
struct page_stat *pstat;
struct page_stat this = {
@@ -888,9 +888,9 @@ static int evsel__process_page_free_event(struct evsel *evsel, struct perf_sampl
};
if (use_pfn)
- page = evsel__intval(evsel, sample, "pfn");
+ page = perf_sample__intval(sample, "pfn");
else
- page = evsel__intval(evsel, sample, "page");
+ page = perf_sample__intval(sample, "page");
nr_page_frees++;
total_page_free_bytes += bytes;
diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c
index d9b9792894a8..dd2ed21596aa 100644
--- a/tools/perf/builtin-kvm.c
+++ b/tools/perf/builtin-kvm.c
@@ -930,7 +930,7 @@ struct vcpu_event_record *per_vcpu_record(struct thread *thread,
return NULL;
}
- vcpu_record->vcpu_id = evsel__intval(sample->evsel, sample, vcpu_id_str(e_machine));
+ vcpu_record->vcpu_id = perf_sample__intval(sample, vcpu_id_str(e_machine));
thread__set_priv(thread, vcpu_record);
}
diff --git a/tools/perf/builtin-kwork.c b/tools/perf/builtin-kwork.c
index d4bb19ed91b5..ab4519287f47 100644
--- a/tools/perf/builtin-kwork.c
+++ b/tools/perf/builtin-kwork.c
@@ -1006,7 +1006,7 @@ static void irq_work_init(struct perf_kwork *kwork,
struct kwork_class *class,
struct kwork_work *work,
enum kwork_trace_type src_type __maybe_unused,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine __maybe_unused)
{
@@ -1014,11 +1014,11 @@ static void irq_work_init(struct perf_kwork *kwork,
work->cpu = sample->cpu;
if (kwork->report == KWORK_REPORT_TOP) {
- work->id = evsel__intval_common(evsel, sample, "common_pid");
+ work->id = perf_sample__intval_common(sample, "common_pid");
work->name = NULL;
} else {
- work->id = evsel__intval(evsel, sample, "irq");
- work->name = evsel__strval(evsel, sample, "name");
+ work->id = perf_sample__intval(sample, "irq");
+ work->name = perf_sample__strval(sample, "name");
}
}
@@ -1144,10 +1144,10 @@ static void softirq_work_init(struct perf_kwork *kwork,
work->cpu = sample->cpu;
if (kwork->report == KWORK_REPORT_TOP) {
- work->id = evsel__intval_common(evsel, sample, "common_pid");
+ work->id = perf_sample__intval_common(sample, "common_pid");
work->name = NULL;
} else {
- num = evsel__intval(evsel, sample, "vec");
+ num = perf_sample__intval(sample, "vec");
work->id = num;
work->name = evsel__softirq_name(evsel, num);
}
@@ -1234,17 +1234,16 @@ static void workqueue_work_init(struct perf_kwork *kwork __maybe_unused,
struct kwork_class *class,
struct kwork_work *work,
enum kwork_trace_type src_type __maybe_unused,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine)
{
char *modp = NULL;
- unsigned long long function_addr = evsel__intval(evsel,
- sample, "function");
+ unsigned long long function_addr = perf_sample__intval(sample, "function");
work->class = class;
work->cpu = sample->cpu;
- work->id = evsel__intval(evsel, sample, "work");
+ work->id = perf_sample__intval(sample, "work");
work->name = function_addr == 0 ? NULL :
machine__resolve_kernel_addr(machine, &function_addr, &modp);
}
@@ -1302,7 +1301,7 @@ static void sched_work_init(struct perf_kwork *kwork __maybe_unused,
struct kwork_class *class,
struct kwork_work *work,
enum kwork_trace_type src_type,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine __maybe_unused)
{
@@ -1310,11 +1309,11 @@ static void sched_work_init(struct perf_kwork *kwork __maybe_unused,
work->cpu = sample->cpu;
if (src_type == KWORK_TRACE_EXIT) {
- work->id = evsel__intval(evsel, sample, "prev_pid");
- work->name = strdup(evsel__strval(evsel, sample, "prev_comm"));
+ work->id = perf_sample__intval(sample, "prev_pid");
+ work->name = strdup(perf_sample__strval(sample, "prev_comm"));
} else if (src_type == KWORK_TRACE_ENTRY) {
- work->id = evsel__intval(evsel, sample, "next_pid");
- work->name = strdup(evsel__strval(evsel, sample, "next_comm"));
+ work->id = perf_sample__intval(sample, "next_pid");
+ work->name = strdup(perf_sample__strval(sample, "next_comm"));
}
}
diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c
index 4d8ddf6391e8..cbf3a39c7837 100644
--- a/tools/perf/builtin-lock.c
+++ b/tools/perf/builtin-lock.c
@@ -563,15 +563,15 @@ static int get_key_by_aggr_mode(u64 *key, u64 addr, struct evsel *evsel,
return get_key_by_aggr_mode_simple(key, addr, sample->tid);
}
-static int report_lock_acquire_event(struct evsel *evsel,
+static int report_lock_acquire_event(struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
struct lock_stat *ls;
struct thread_stat *ts;
struct lock_seq_stat *seq;
- const char *name = evsel__strval(evsel, sample, "name");
- u64 addr = evsel__intval(evsel, sample, "lockdep_addr");
- int flag = evsel__intval(evsel, sample, "flags");
+ const char *name = perf_sample__strval(sample, "name");
+ u64 addr = perf_sample__intval(sample, "lockdep_addr");
+ int flag = perf_sample__intval(sample, "flags");
u64 key;
int ret;
@@ -638,15 +638,15 @@ static int report_lock_acquire_event(struct evsel *evsel,
return 0;
}
-static int report_lock_acquired_event(struct evsel *evsel,
+static int report_lock_acquired_event(struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
struct lock_stat *ls;
struct thread_stat *ts;
struct lock_seq_stat *seq;
u64 contended_term;
- const char *name = evsel__strval(evsel, sample, "name");
- u64 addr = evsel__intval(evsel, sample, "lockdep_addr");
+ const char *name = perf_sample__strval(sample, "name");
+ u64 addr = perf_sample__intval(sample, "lockdep_addr");
u64 key;
int ret;
@@ -704,14 +704,14 @@ static int report_lock_acquired_event(struct evsel *evsel,
return 0;
}
-static int report_lock_contended_event(struct evsel *evsel,
+static int report_lock_contended_event(struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
struct lock_stat *ls;
struct thread_stat *ts;
struct lock_seq_stat *seq;
- const char *name = evsel__strval(evsel, sample, "name");
- u64 addr = evsel__intval(evsel, sample, "lockdep_addr");
+ const char *name = perf_sample__strval(sample, "name");
+ u64 addr = perf_sample__intval(sample, "lockdep_addr");
u64 key;
int ret;
@@ -762,14 +762,14 @@ static int report_lock_contended_event(struct evsel *evsel,
return 0;
}
-static int report_lock_release_event(struct evsel *evsel,
+static int report_lock_release_event(struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
struct lock_stat *ls;
struct thread_stat *ts;
struct lock_seq_stat *seq;
- const char *name = evsel__strval(evsel, sample, "name");
- u64 addr = evsel__intval(evsel, sample, "lockdep_addr");
+ const char *name = perf_sample__strval(sample, "name");
+ u64 addr = perf_sample__intval(sample, "lockdep_addr");
u64 key;
int ret;
@@ -969,8 +969,8 @@ static int report_lock_contention_begin_event(struct evsel *evsel,
struct lock_stat *ls;
struct thread_stat *ts;
struct lock_seq_stat *seq;
- u64 addr = evsel__intval(evsel, sample, "lock_addr");
- unsigned int flags = evsel__intval(evsel, sample, "flags");
+ u64 addr = perf_sample__intval(sample, "lock_addr");
+ unsigned int flags = perf_sample__intval(sample, "flags");
u64 key;
int i, ret;
static bool kmap_loaded;
@@ -1134,7 +1134,7 @@ static int report_lock_contention_end_event(struct evsel *evsel,
struct thread_stat *ts;
struct lock_seq_stat *seq;
u64 contended_term;
- u64 addr = evsel__intval(evsel, sample, "lock_addr");
+ u64 addr = perf_sample__intval(sample, "lock_addr");
u64 key;
int ret;
diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
index 2950c5c3026e..01228f630121 100644
--- a/tools/perf/builtin-sched.c
+++ b/tools/perf/builtin-sched.c
@@ -829,8 +829,8 @@ replay_wakeup_event(struct perf_sched *sched,
struct evsel *evsel, struct perf_sample *sample,
struct machine *machine __maybe_unused)
{
- const char *comm = evsel__strval(evsel, sample, "comm");
- const u32 pid = evsel__intval(evsel, sample, "pid");
+ const char *comm = perf_sample__strval(sample, "comm");
+ const u32 pid = perf_sample__intval(sample, "pid");
struct task_desc *waker, *wakee;
if (verbose > 0) {
@@ -851,10 +851,10 @@ static int replay_switch_event(struct perf_sched *sched,
struct perf_sample *sample,
struct machine *machine __maybe_unused)
{
- const char *prev_comm = evsel__strval(evsel, sample, "prev_comm"),
- *next_comm = evsel__strval(evsel, sample, "next_comm");
- const u32 prev_pid = evsel__intval(evsel, sample, "prev_pid"),
- next_pid = evsel__intval(evsel, sample, "next_pid");
+ const char *prev_comm = perf_sample__strval(sample, "prev_comm"),
+ *next_comm = perf_sample__strval(sample, "next_comm");
+ const u32 prev_pid = perf_sample__intval(sample, "prev_pid"),
+ next_pid = perf_sample__intval(sample, "next_pid");
struct task_desc *prev, __maybe_unused *next;
u64 timestamp0, timestamp = sample->time;
int cpu = sample->cpu;
@@ -1134,13 +1134,13 @@ static void free_work_atoms(struct work_atoms *atoms)
}
static int latency_switch_event(struct perf_sched *sched,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine)
{
- const u32 prev_pid = evsel__intval(evsel, sample, "prev_pid"),
- next_pid = evsel__intval(evsel, sample, "next_pid");
- const char prev_state = evsel__taskstate(evsel, sample, "prev_state");
+ const u32 prev_pid = perf_sample__intval(sample, "prev_pid"),
+ next_pid = perf_sample__intval(sample, "next_pid");
+ const char prev_state = perf_sample__taskstate(sample, "prev_state");
struct work_atoms *out_events, *in_events;
struct thread *sched_out, *sched_in;
u64 timestamp0, timestamp = sample->time;
@@ -1204,12 +1204,12 @@ static int latency_switch_event(struct perf_sched *sched,
}
static int latency_runtime_event(struct perf_sched *sched,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine)
{
- const u32 pid = evsel__intval(evsel, sample, "pid");
- const u64 runtime = evsel__intval(evsel, sample, "runtime");
+ const u32 pid = perf_sample__intval(sample, "pid");
+ const u64 runtime = perf_sample__intval(sample, "runtime");
struct thread *thread = machine__findnew_thread(machine, -1, pid);
struct work_atoms *atoms = thread_atoms_search(&sched->atom_root, thread, &sched->cmp_pid);
u64 timestamp = sample->time;
@@ -1239,11 +1239,11 @@ static int latency_runtime_event(struct perf_sched *sched,
}
static int latency_wakeup_event(struct perf_sched *sched,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine)
{
- const u32 pid = evsel__intval(evsel, sample, "pid");
+ const u32 pid = perf_sample__intval(sample, "pid");
struct work_atoms *atoms;
struct work_atom *atom;
struct thread *wakee;
@@ -1300,11 +1300,11 @@ static int latency_wakeup_event(struct perf_sched *sched,
}
static int latency_migrate_task_event(struct perf_sched *sched,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine)
{
- const u32 pid = evsel__intval(evsel, sample, "pid");
+ const u32 pid = perf_sample__intval(sample, "pid");
u64 timestamp = sample->time;
struct work_atoms *atoms;
struct work_atom *atom;
@@ -1626,11 +1626,11 @@ static void print_sched_map(struct perf_sched *sched, struct perf_cpu this_cpu,
}
}
-static int map_switch_event(struct perf_sched *sched, struct evsel *evsel,
+static int map_switch_event(struct perf_sched *sched, struct evsel *evsel __maybe_unused,
struct perf_sample *sample, struct machine *machine)
{
- const u32 next_pid = evsel__intval(evsel, sample, "next_pid");
- const u32 prev_pid = evsel__intval(evsel, sample, "prev_pid");
+ const u32 next_pid = perf_sample__intval(sample, "next_pid");
+ const u32 prev_pid = perf_sample__intval(sample, "prev_pid");
struct thread *sched_in, *sched_out;
struct thread_runtime *tr;
int new_shortname;
@@ -1797,8 +1797,8 @@ static int process_sched_switch_event(const struct perf_tool *tool,
{
struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
int this_cpu = sample->cpu, err = 0;
- u32 prev_pid = evsel__intval(evsel, sample, "prev_pid"),
- next_pid = evsel__intval(evsel, sample, "next_pid");
+ u32 prev_pid = perf_sample__intval(sample, "prev_pid"),
+ next_pid = perf_sample__intval(sample, "next_pid");
if (sched->curr_pid[this_cpu] != (u32)-1) {
/*
@@ -2068,12 +2068,11 @@ static char *timehist_get_commstr(struct thread *thread)
/* prio field format: xxx or xxx->yyy */
#define MAX_PRIO_STR_LEN 8
-static char *timehist_get_priostr(struct evsel *evsel,
- struct thread *thread,
+static char *timehist_get_priostr(struct thread *thread,
struct perf_sample *sample)
{
static char prio_str[16];
- int prev_prio = (int)evsel__intval(evsel, sample, "prev_prio");
+ int prev_prio = (int)perf_sample__intval(sample, "prev_prio");
struct thread_runtime *tr = thread__priv(thread);
if (tr->prio != prev_prio && tr->prio != -1)
@@ -2161,15 +2160,14 @@ static void timehist_header(struct perf_sched *sched)
}
static void timehist_print_sample(struct perf_sched *sched,
- struct evsel *evsel,
struct perf_sample *sample,
struct addr_location *al,
struct thread *thread,
u64 t, const char state)
{
struct thread_runtime *tr = thread__priv(thread);
- const char *next_comm = evsel__strval(evsel, sample, "next_comm");
- const u32 next_pid = evsel__intval(evsel, sample, "next_pid");
+ const char *next_comm = perf_sample__strval(sample, "next_comm");
+ const u32 next_pid = perf_sample__intval(sample, "next_pid");
u32 max_cpus = sched->max_cpu.cpu + 1;
char tstr[64];
char nstr[30];
@@ -2198,14 +2196,15 @@ static void timehist_print_sample(struct perf_sched *sched,
}
if (!thread__comm_set(thread)) {
- const char *prev_comm = evsel__strval(evsel, sample, "prev_comm");
- thread__set_comm(thread, prev_comm, sample->time);
+ const char *prev_comm = perf_sample__strval(sample, "prev_comm");
+
+ thread__set_comm(thread, prev_comm, sample->time);
}
printf(" %-*s ", comm_width, timehist_get_commstr(thread));
if (sched->show_prio)
- printf(" %-*s ", MAX_PRIO_STR_LEN, timehist_get_priostr(evsel, thread, sample));
+ printf(" %-*s ", MAX_PRIO_STR_LEN, timehist_get_priostr(thread, sample));
wait_time = tr->dt_sleep + tr->dt_iowait + tr->dt_preempt;
print_sched_time(wait_time, 6);
@@ -2319,7 +2318,7 @@ static bool is_idle_sample(struct perf_sample *sample,
{
/* pid 0 == swapper == idle task */
if (evsel__name_is(evsel, "sched:sched_switch"))
- return evsel__intval(evsel, sample, "prev_pid") == 0;
+ return perf_sample__intval(sample, "prev_pid") == 0;
return sample->pid == 0;
}
@@ -2539,7 +2538,7 @@ static struct thread *timehist_get_thread(struct perf_sched *sched,
itr->last_thread = thread__get(thread);
/* copy task callchain when entering to idle */
- if (evsel__intval(evsel, sample, "next_pid") == 0)
+ if (perf_sample__intval(sample, "next_pid") == 0)
save_idle_callchain(sched, itr, sample);
}
}
@@ -2572,7 +2571,7 @@ static bool timehist_skip_sample(struct perf_sched *sched,
if (tr && tr->prio != -1)
prio = tr->prio;
else if (evsel__name_is(evsel, "sched:sched_switch"))
- prio = evsel__intval(evsel, sample, "prev_prio");
+ prio = perf_sample__intval(sample, "prev_prio");
if (prio != -1 && !test_bit(prio, sched->prio_bitmap)) {
rc = true;
@@ -2583,8 +2582,8 @@ static bool timehist_skip_sample(struct perf_sched *sched,
if (sched->idle_hist) {
if (!evsel__name_is(evsel, "sched:sched_switch"))
rc = true;
- else if (evsel__intval(evsel, sample, "prev_pid") != 0 &&
- evsel__intval(evsel, sample, "next_pid") != 0)
+ else if (perf_sample__intval(sample, "prev_pid") != 0 &&
+ perf_sample__intval(sample, "next_pid") != 0)
rc = true;
}
@@ -2647,7 +2646,7 @@ static int timehist_sched_wakeup_event(const struct perf_tool *tool,
struct thread *thread;
struct thread_runtime *tr = NULL;
/* want pid of awakened task not pid in sample */
- const u32 pid = evsel__intval(evsel, sample, "pid");
+ const u32 pid = perf_sample__intval(sample, "pid");
thread = machine__findnew_thread(machine, 0, pid);
if (thread == NULL)
@@ -2686,8 +2685,8 @@ static void timehist_print_migration_event(struct perf_sched *sched,
return;
max_cpus = sched->max_cpu.cpu + 1;
- ocpu = evsel__intval(evsel, sample, "orig_cpu");
- dcpu = evsel__intval(evsel, sample, "dest_cpu");
+ ocpu = perf_sample__intval(sample, "orig_cpu");
+ dcpu = perf_sample__intval(sample, "dest_cpu");
thread = machine__findnew_thread(machine, sample->pid, sample->tid);
if (thread == NULL)
@@ -2736,7 +2735,7 @@ static int timehist_migrate_task_event(const struct perf_tool *tool,
struct thread *thread;
struct thread_runtime *tr = NULL;
/* want pid of migrated task not pid in sample */
- const u32 pid = evsel__intval(evsel, sample, "pid");
+ const u32 pid = perf_sample__intval(sample, "pid");
thread = machine__findnew_thread(machine, 0, pid);
if (thread == NULL)
@@ -2761,14 +2760,13 @@ static int timehist_migrate_task_event(const struct perf_tool *tool,
return 0;
}
-static void timehist_update_task_prio(struct evsel *evsel,
- struct perf_sample *sample,
+static void timehist_update_task_prio(struct perf_sample *sample,
struct machine *machine)
{
struct thread *thread;
struct thread_runtime *tr = NULL;
- const u32 next_pid = evsel__intval(evsel, sample, "next_pid");
- const u32 next_prio = evsel__intval(evsel, sample, "next_prio");
+ const u32 next_pid = perf_sample__intval(sample, "next_pid");
+ const u32 next_prio = perf_sample__intval(sample, "next_prio");
if (next_pid == 0)
thread = get_idle_thread(sample->cpu);
@@ -2798,7 +2796,7 @@ static int timehist_sched_change_event(const struct perf_tool *tool,
struct thread_runtime *tr = NULL;
u64 tprev, t = sample->time;
int rc = 0;
- const char state = evsel__taskstate(evsel, sample, "prev_state");
+ const char state = perf_sample__taskstate(sample, "prev_state");
addr_location__init(&al);
if (machine__resolve(machine, &al, sample) < 0) {
@@ -2809,7 +2807,7 @@ static int timehist_sched_change_event(const struct perf_tool *tool,
}
if (sched->show_prio || sched->prio_str)
- timehist_update_task_prio(evsel, sample, machine);
+ timehist_update_task_prio(sample, machine);
thread = timehist_get_thread(sched, sample, machine, evsel);
if (thread == NULL) {
@@ -2891,7 +2889,7 @@ static int timehist_sched_change_event(const struct perf_tool *tool,
}
if (!sched->summary_only)
- timehist_print_sample(sched, evsel, sample, &al, thread, t, state);
+ timehist_print_sample(sched, sample, &al, thread, t, state);
}
out:
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c
index 8692d11ccd29..034e0ba3f9cb 100644
--- a/tools/perf/builtin-timechart.c
+++ b/tools/perf/builtin-timechart.c
@@ -597,12 +597,12 @@ static int process_sample_event(const struct perf_tool *tool,
static int
process_sample_cpu_idle(struct timechart *tchart __maybe_unused,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace __maybe_unused)
{
- u32 state = evsel__intval(evsel, sample, "state");
- u32 cpu_id = evsel__intval(evsel, sample, "cpu_id");
+ u32 state = perf_sample__intval(sample, "state");
+ u32 cpu_id = perf_sample__intval(sample, "cpu_id");
if (state == (u32)PWR_EVENT_EXIT)
c_state_end(tchart, cpu_id, sample->time);
@@ -613,12 +613,12 @@ process_sample_cpu_idle(struct timechart *tchart __maybe_unused,
static int
process_sample_cpu_frequency(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace __maybe_unused)
{
- u32 state = evsel__intval(evsel, sample, "state");
- u32 cpu_id = evsel__intval(evsel, sample, "cpu_id");
+ u32 state = perf_sample__intval(sample, "state");
+ u32 cpu_id = perf_sample__intval(sample, "cpu_id");
p_state_change(tchart, cpu_id, sample->time, state);
return 0;
@@ -626,13 +626,13 @@ process_sample_cpu_frequency(struct timechart *tchart,
static int
process_sample_sched_wakeup(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace)
{
- u8 flags = evsel__intval(evsel, sample, "common_flags");
- int waker = evsel__intval(evsel, sample, "common_pid");
- int wakee = evsel__intval(evsel, sample, "pid");
+ u8 flags = perf_sample__intval(sample, "common_flags");
+ int waker = perf_sample__intval(sample, "common_pid");
+ int wakee = perf_sample__intval(sample, "pid");
sched_wakeup(tchart, sample->cpu, sample->time, waker, wakee, flags, backtrace);
return 0;
@@ -640,13 +640,13 @@ process_sample_sched_wakeup(struct timechart *tchart,
static int
process_sample_sched_switch(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace)
{
- int prev_pid = evsel__intval(evsel, sample, "prev_pid");
- int next_pid = evsel__intval(evsel, sample, "next_pid");
- u64 prev_state = evsel__intval(evsel, sample, "prev_state");
+ int prev_pid = perf_sample__intval(sample, "prev_pid");
+ int next_pid = perf_sample__intval(sample, "next_pid");
+ u64 prev_state = perf_sample__intval(sample, "prev_state");
sched_switch(tchart, sample->cpu, sample->time, prev_pid, next_pid,
prev_state, backtrace);
@@ -656,12 +656,12 @@ process_sample_sched_switch(struct timechart *tchart,
#ifdef SUPPORT_OLD_POWER_EVENTS
static int
process_sample_power_start(struct timechart *tchart __maybe_unused,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace __maybe_unused)
{
- u64 cpu_id = evsel__intval(evsel, sample, "cpu_id");
- u64 value = evsel__intval(evsel, sample, "value");
+ u64 cpu_id = perf_sample__intval(sample, "cpu_id");
+ u64 value = perf_sample__intval(sample, "value");
c_state_start(cpu_id, sample->time, value);
return 0;
@@ -679,12 +679,12 @@ process_sample_power_end(struct timechart *tchart,
static int
process_sample_power_frequency(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace __maybe_unused)
{
- u64 cpu_id = evsel__intval(evsel, sample, "cpu_id");
- u64 value = evsel__intval(evsel, sample, "value");
+ u64 cpu_id = perf_sample__intval(sample, "cpu_id");
+ u64 value = perf_sample__intval(sample, "value");
p_state_change(tchart, cpu_id, sample->time, value);
return 0;
@@ -849,120 +849,120 @@ static int pid_end_io_sample(struct timechart *tchart, int pid, int type,
static int
process_enter_read(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
- long fd = evsel__intval(evsel, sample, "fd");
+ long fd = perf_sample__intval(sample, "fd");
return pid_begin_io_sample(tchart, sample->tid, IOTYPE_READ,
sample->time, fd);
}
static int
process_exit_read(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
- long ret = evsel__intval(evsel, sample, "ret");
+ long ret = perf_sample__intval(sample, "ret");
return pid_end_io_sample(tchart, sample->tid, IOTYPE_READ,
sample->time, ret);
}
static int
process_enter_write(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
- long fd = evsel__intval(evsel, sample, "fd");
+ long fd = perf_sample__intval(sample, "fd");
return pid_begin_io_sample(tchart, sample->tid, IOTYPE_WRITE,
sample->time, fd);
}
static int
process_exit_write(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
- long ret = evsel__intval(evsel, sample, "ret");
+ long ret = perf_sample__intval(sample, "ret");
return pid_end_io_sample(tchart, sample->tid, IOTYPE_WRITE,
sample->time, ret);
}
static int
process_enter_sync(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
- long fd = evsel__intval(evsel, sample, "fd");
+ long fd = perf_sample__intval(sample, "fd");
return pid_begin_io_sample(tchart, sample->tid, IOTYPE_SYNC,
sample->time, fd);
}
static int
process_exit_sync(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
- long ret = evsel__intval(evsel, sample, "ret");
+ long ret = perf_sample__intval(sample, "ret");
return pid_end_io_sample(tchart, sample->tid, IOTYPE_SYNC,
sample->time, ret);
}
static int
process_enter_tx(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
- long fd = evsel__intval(evsel, sample, "fd");
+ long fd = perf_sample__intval(sample, "fd");
return pid_begin_io_sample(tchart, sample->tid, IOTYPE_TX,
sample->time, fd);
}
static int
process_exit_tx(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
- long ret = evsel__intval(evsel, sample, "ret");
+ long ret = perf_sample__intval(sample, "ret");
return pid_end_io_sample(tchart, sample->tid, IOTYPE_TX,
sample->time, ret);
}
static int
process_enter_rx(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
- long fd = evsel__intval(evsel, sample, "fd");
+ long fd = perf_sample__intval(sample, "fd");
return pid_begin_io_sample(tchart, sample->tid, IOTYPE_RX,
sample->time, fd);
}
static int
process_exit_rx(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
- long ret = evsel__intval(evsel, sample, "ret");
+ long ret = perf_sample__intval(sample, "ret");
return pid_end_io_sample(tchart, sample->tid, IOTYPE_RX,
sample->time, ret);
}
static int
process_enter_poll(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
- long fd = evsel__intval(evsel, sample, "fd");
+ long fd = perf_sample__intval(sample, "fd");
return pid_begin_io_sample(tchart, sample->tid, IOTYPE_POLL,
sample->time, fd);
}
static int
process_exit_poll(struct timechart *tchart,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
- long ret = evsel__intval(evsel, sample, "ret");
+ long ret = perf_sample__intval(sample, "ret");
return pid_end_io_sample(tchart, sample->tid, IOTYPE_POLL,
sample->time, ret);
}
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index 1aafa77f72cf..e8aabd7a72fe 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -3094,7 +3094,7 @@ errno_print: {
return err;
}
-static int trace__vfs_getname(struct trace *trace, struct evsel *evsel,
+static int trace__vfs_getname(struct trace *trace, struct evsel *evsel __maybe_unused,
union perf_event *event __maybe_unused,
struct perf_sample *sample)
{
@@ -3103,7 +3103,7 @@ static int trace__vfs_getname(struct trace *trace, struct evsel *evsel,
size_t filename_len, entry_str_len, to_move;
ssize_t remaining_space;
char *pos;
- const char *filename = evsel__rawptr(evsel, sample, "pathname");
+ const char *filename = perf_sample__strval(sample, "pathname");
if (!thread)
goto out;
@@ -3159,7 +3159,7 @@ static int trace__sched_stat_runtime(struct trace *trace, struct evsel *evsel,
union perf_event *event __maybe_unused,
struct perf_sample *sample)
{
- u64 runtime = evsel__intval(evsel, sample, "runtime");
+ u64 runtime = perf_sample__intval(sample, "runtime");
double runtime_ms = (double)runtime / NSEC_PER_MSEC;
struct thread *thread = machine__findnew_thread(trace->host,
sample->pid,
@@ -3178,10 +3178,10 @@ static int trace__sched_stat_runtime(struct trace *trace, struct evsel *evsel,
out_dump:
fprintf(trace->output, "%s: comm=%s,pid=%u,runtime=%" PRIu64 ",vruntime=%" PRIu64 ")\n",
evsel->name,
- evsel__strval(evsel, sample, "comm"),
- (pid_t)evsel__intval(evsel, sample, "pid"),
+ perf_sample__strval(sample, "comm"),
+ (pid_t)perf_sample__intval(sample, "pid"),
runtime,
- evsel__intval(evsel, sample, "vruntime"));
+ perf_sample__intval(sample, "vruntime"));
goto out_put;
}
diff --git a/tools/perf/tests/openat-syscall-tp-fields.c b/tools/perf/tests/openat-syscall-tp-fields.c
index 5523cf4e9321..9ff8caff98c3 100644
--- a/tools/perf/tests/openat-syscall-tp-fields.c
+++ b/tools/perf/tests/openat-syscall-tp-fields.c
@@ -118,7 +118,7 @@ static int test__syscall_openat_tp_fields(struct test_suite *test __maybe_unused
goto out_delete_evlist;
}
- tp_flags = evsel__intval(evsel, &sample, "flags");
+ tp_flags = perf_sample__intval(&sample, "flags");
perf_sample__exit(&sample);
/* C library wrapper may set additional flags,
access mode must be unchanged */
diff --git a/tools/perf/tests/switch-tracking.c b/tools/perf/tests/switch-tracking.c
index 72a8289e846d..22b0302252db 100644
--- a/tools/perf/tests/switch-tracking.c
+++ b/tools/perf/tests/switch-tracking.c
@@ -140,8 +140,8 @@ static int process_sample_event(struct evlist *evlist,
evsel = evlist__id2evsel(evlist, sample.id);
if (evsel == switch_tracking->switch_evsel) {
- next_tid = evsel__intval(evsel, &sample, "next_pid");
- prev_tid = evsel__intval(evsel, &sample, "prev_pid");
+ next_tid = perf_sample__intval(&sample, "next_pid");
+ prev_tid = perf_sample__intval(&sample, "prev_pid");
cpu = sample.cpu;
pr_debug3("sched_switch: cpu: %d prev_tid %d next_tid %d\n",
cpu, prev_tid, next_tid);
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index 2ee87fd84d3e..ee30e15af054 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -3687,9 +3687,9 @@ struct tep_format_field *evsel__common_field(struct evsel *evsel, const char *na
return tp_format ? tep_find_common_field(tp_format, name) : NULL;
}
-void *evsel__rawptr(struct evsel *evsel, struct perf_sample *sample, const char *name)
+void *perf_sample__rawptr(struct perf_sample *sample, const char *name)
{
- struct tep_format_field *field = evsel__field(evsel, name);
+ struct tep_format_field *field = evsel__field(sample->evsel, name);
int offset;
if (!field)
@@ -3746,21 +3746,21 @@ u64 format_field__intval(struct tep_format_field *field, struct perf_sample *sam
return 0;
}
-u64 evsel__intval(struct evsel *evsel, struct perf_sample *sample, const char *name)
+u64 perf_sample__intval(struct perf_sample *sample, const char *name)
{
- struct tep_format_field *field = evsel__field(evsel, name);
+ struct tep_format_field *field = evsel__field(sample->evsel, name);
- return field ? format_field__intval(field, sample, evsel->needs_swap) : 0;
+ return field ? format_field__intval(field, sample, sample->evsel->needs_swap) : 0;
}
-u64 evsel__intval_common(struct evsel *evsel, struct perf_sample *sample, const char *name)
+u64 perf_sample__intval_common(struct perf_sample *sample, const char *name)
{
- struct tep_format_field *field = evsel__common_field(evsel, name);
+ struct tep_format_field *field = evsel__common_field(sample->evsel, name);
- return field ? format_field__intval(field, sample, evsel->needs_swap) : 0;
+ return field ? format_field__intval(field, sample, sample->evsel->needs_swap) : 0;
}
-char evsel__taskstate(struct evsel *evsel, struct perf_sample *sample, const char *name)
+char perf_sample__taskstate(struct perf_sample *sample, const char *name)
{
static struct tep_format_field *prev_state_field;
static const char *states;
@@ -3769,7 +3769,7 @@ char evsel__taskstate(struct evsel *evsel, struct perf_sample *sample, const cha
unsigned int bit;
char state = '?'; /* '?' denotes unknown task state */
- field = evsel__field(evsel, name);
+ field = evsel__field(sample->evsel, name);
if (!field)
return state;
@@ -3788,7 +3788,7 @@ char evsel__taskstate(struct evsel *evsel, struct perf_sample *sample, const cha
*
* We can change this if we have a good reason in the future.
*/
- val = evsel__intval(evsel, sample, name);
+ val = perf_sample__intval(sample, name);
bit = val ? ffs(val) : 0;
state = (!bit || bit > strlen(states)) ? 'R' : states[bit-1];
return state;
diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
index 339b5c08a33d..927e5b4756cc 100644
--- a/tools/perf/util/evsel.h
+++ b/tools/perf/util/evsel.h
@@ -371,14 +371,14 @@ bool evsel__precise_ip_fallback(struct evsel *evsel);
struct perf_sample;
#ifdef HAVE_LIBTRACEEVENT
-void *evsel__rawptr(struct evsel *evsel, struct perf_sample *sample, const char *name);
-u64 evsel__intval(struct evsel *evsel, struct perf_sample *sample, const char *name);
-u64 evsel__intval_common(struct evsel *evsel, struct perf_sample *sample, const char *name);
-char evsel__taskstate(struct evsel *evsel, struct perf_sample *sample, const char *name);
+void *perf_sample__rawptr(struct perf_sample *sample, const char *name);
+u64 perf_sample__intval(struct perf_sample *sample, const char *name);
+u64 perf_sample__intval_common(struct perf_sample *sample, const char *name);
+char perf_sample__taskstate(struct perf_sample *sample, const char *name);
-static inline char *evsel__strval(struct evsel *evsel, struct perf_sample *sample, const char *name)
+static inline char *perf_sample__strval(struct perf_sample *sample, const char *name)
{
- return evsel__rawptr(evsel, sample, name);
+ return perf_sample__rawptr(sample, name);
}
#endif
diff --git a/tools/perf/util/intel-pt.c b/tools/perf/util/intel-pt.c
index fc9eec8b54b8..dab23a96b1d8 100644
--- a/tools/perf/util/intel-pt.c
+++ b/tools/perf/util/intel-pt.c
@@ -3426,7 +3426,7 @@ static int intel_pt_process_switch(struct intel_pt *pt,
if (evsel != pt->switch_evsel)
return 0;
- tid = evsel__intval(evsel, sample, "next_pid");
+ tid = perf_sample__intval(sample, "next_pid");
cpu = sample->cpu;
intel_pt_log("sched_switch: cpu %d tid %d time %"PRIu64" tsc %#"PRIx64"\n",
diff --git a/tools/perf/util/kvm-stat-arch/kvm-stat-arm64.c b/tools/perf/util/kvm-stat-arch/kvm-stat-arm64.c
index 1e76906f719c..018b0db0e6e7 100644
--- a/tools/perf/util/kvm-stat-arch/kvm-stat-arm64.c
+++ b/tools/perf/util/kvm-stat-arch/kvm-stat-arm64.c
@@ -20,10 +20,8 @@ static const char * const __kvm_events_tp[] = {
static void event_get_key(struct perf_sample *sample,
struct event_key *key)
{
- struct evsel *evsel = sample->evsel;
-
key->info = 0;
- key->key = evsel__intval(evsel, sample, kvm_exit_reason(EM_AARCH64));
+ key->key = perf_sample__intval(sample, kvm_exit_reason(EM_AARCH64));
key->exit_reasons = arm64_exit_reasons;
/*
@@ -32,7 +30,7 @@ static void event_get_key(struct perf_sample *sample,
* properly decode event's est_ec.
*/
if (key->key == ARM_EXCEPTION_TRAP) {
- key->key = evsel__intval(evsel, sample, kvm_trap_exit_reason);
+ key->key = perf_sample__intval(sample, kvm_trap_exit_reason);
key->exit_reasons = arm64_trap_exit_reasons;
}
}
diff --git a/tools/perf/util/kvm-stat-arch/kvm-stat-loongarch.c b/tools/perf/util/kvm-stat-arch/kvm-stat-loongarch.c
index 9d6265290f6d..a04cd09e3361 100644
--- a/tools/perf/util/kvm-stat-arch/kvm-stat-loongarch.c
+++ b/tools/perf/util/kvm-stat-arch/kvm-stat-loongarch.c
@@ -78,7 +78,7 @@ static void event_gspr_get_key(struct perf_sample *sample, struct event_key *key
unsigned int insn;
key->key = LOONGARCH_EXCEPTION_OTHERS;
- insn = evsel__intval(sample->evsel, sample, "inst_word");
+ insn = perf_sample__intval(sample, "inst_word");
switch (insn >> 24) {
case 0:
diff --git a/tools/perf/util/kvm-stat-arch/kvm-stat-powerpc.c b/tools/perf/util/kvm-stat-arch/kvm-stat-powerpc.c
index 5158d7e88ee6..96d9c4ae0209 100644
--- a/tools/perf/util/kvm-stat-arch/kvm-stat-powerpc.c
+++ b/tools/perf/util/kvm-stat-arch/kvm-stat-powerpc.c
@@ -32,7 +32,7 @@ static void hcall_event_get_key(struct perf_sample *sample,
struct event_key *key)
{
key->info = 0;
- key->key = evsel__intval(sample->evsel, sample, "req");
+ key->key = perf_sample__intval(sample, "req");
}
static const char *get_hcall_exit_reason(u64 exit_code)
diff --git a/tools/perf/util/kvm-stat-arch/kvm-stat-riscv.c b/tools/perf/util/kvm-stat-arch/kvm-stat-riscv.c
index e8db8b4f8e2e..967bba261a47 100644
--- a/tools/perf/util/kvm-stat-arch/kvm-stat-riscv.c
+++ b/tools/perf/util/kvm-stat-arch/kvm-stat-riscv.c
@@ -26,8 +26,7 @@ static void event_get_key(struct perf_sample *sample,
int xlen = 64; // TODO: 32-bit support.
key->info = 0;
- key->key = evsel__intval(sample->evsel, sample,
- kvm_exit_reason(EM_RISCV)) & ~CAUSE_IRQ_FLAG(xlen);
+ key->key = perf_sample__intval(sample, kvm_exit_reason(EM_RISCV)) & ~CAUSE_IRQ_FLAG(xlen);
key->exit_reasons = riscv_exit_reasons;
}
diff --git a/tools/perf/util/kvm-stat-arch/kvm-stat-s390.c b/tools/perf/util/kvm-stat-arch/kvm-stat-s390.c
index 158372ba0205..4771fc69fa39 100644
--- a/tools/perf/util/kvm-stat-arch/kvm-stat-s390.c
+++ b/tools/perf/util/kvm-stat-arch/kvm-stat-s390.c
@@ -23,7 +23,7 @@ static void event_icpt_insn_get_key(struct perf_sample *sample,
{
u64 insn;
- insn = evsel__intval(sample->evsel, sample, "instruction");
+ insn = perf_sample__intval(sample, "instruction");
key->key = icpt_insn_decoder(insn);
key->exit_reasons = sie_icpt_insn_codes;
}
@@ -31,21 +31,21 @@ static void event_icpt_insn_get_key(struct perf_sample *sample,
static void event_sigp_get_key(struct perf_sample *sample,
struct event_key *key)
{
- key->key = evsel__intval(sample->evsel, sample, "order_code");
+ key->key = perf_sample__intval(sample, "order_code");
key->exit_reasons = sie_sigp_order_codes;
}
static void event_diag_get_key(struct perf_sample *sample,
struct event_key *key)
{
- key->key = evsel__intval(sample->evsel, sample, "code");
+ key->key = perf_sample__intval(sample, "code");
key->exit_reasons = sie_diagnose_codes;
}
static void event_icpt_prog_get_key(struct perf_sample *sample,
struct event_key *key)
{
- key->key = evsel__intval(sample->evsel, sample, "code");
+ key->key = perf_sample__intval(sample, "code");
key->exit_reasons = sie_icpt_prog_codes;
}
diff --git a/tools/perf/util/kvm-stat-arch/kvm-stat-x86.c b/tools/perf/util/kvm-stat-arch/kvm-stat-x86.c
index 0ce543d82850..788d216f0852 100644
--- a/tools/perf/util/kvm-stat-arch/kvm-stat-x86.c
+++ b/tools/perf/util/kvm-stat-arch/kvm-stat-x86.c
@@ -27,8 +27,8 @@ static const struct kvm_events_ops exit_events = {
static void mmio_event_get_key(struct perf_sample *sample,
struct event_key *key)
{
- key->key = evsel__intval(sample->evsel, sample, "gpa");
- key->info = evsel__intval(sample->evsel, sample, "type");
+ key->key = perf_sample__intval(sample, "gpa");
+ key->info = perf_sample__intval(sample, "type");
}
#define KVM_TRACE_MMIO_READ_UNSATISFIED 0
@@ -43,7 +43,7 @@ static bool mmio_event_begin(struct perf_sample *sample, struct event_key *key)
/* MMIO write begin event in kernel. */
if (evsel__name_is(sample->evsel, "kvm:kvm_mmio") &&
- evsel__intval(sample->evsel, sample, "type") == KVM_TRACE_MMIO_WRITE) {
+ perf_sample__intval(sample, "type") == KVM_TRACE_MMIO_WRITE) {
mmio_event_get_key(sample, key);
return true;
}
@@ -59,7 +59,7 @@ static bool mmio_event_end(struct perf_sample *sample, struct event_key *key)
/* MMIO read end event in kernel.*/
if (evsel__name_is(sample->evsel, "kvm:kvm_mmio") &&
- evsel__intval(sample->evsel, sample, "type") == KVM_TRACE_MMIO_READ) {
+ perf_sample__intval(sample, "type") == KVM_TRACE_MMIO_READ) {
mmio_event_get_key(sample, key);
return true;
}
@@ -87,8 +87,8 @@ static const struct kvm_events_ops mmio_events = {
static void ioport_event_get_key(struct perf_sample *sample,
struct event_key *key)
{
- key->key = evsel__intval(sample->evsel, sample, "port");
- key->info = evsel__intval(sample->evsel, sample, "rw");
+ key->key = perf_sample__intval(sample, "port");
+ key->info = perf_sample__intval(sample, "rw");
}
static bool ioport_event_begin(struct perf_sample *sample,
@@ -126,8 +126,8 @@ static const struct kvm_events_ops ioport_events = {
/* The time of emulation msr is from kvm_msr to kvm_entry. */
static void msr_event_get_key(struct perf_sample *sample, struct event_key *key)
{
- key->key = evsel__intval(sample->evsel, sample, "ecx");
- key->info = evsel__intval(sample->evsel, sample, "write");
+ key->key = perf_sample__intval(sample, "ecx");
+ key->info = perf_sample__intval(sample, "write");
}
static bool msr_event_begin(struct perf_sample *sample, struct event_key *key)
diff --git a/tools/perf/util/kvm-stat.c b/tools/perf/util/kvm-stat.c
index f17a6132958d..755ab659a05c 100644
--- a/tools/perf/util/kvm-stat.c
+++ b/tools/perf/util/kvm-stat.c
@@ -17,7 +17,7 @@ void exit_event_get_key(struct perf_sample *sample,
uint16_t e_machine = evsel__e_machine(sample->evsel, /*e_flags=*/NULL);
key->info = 0;
- key->key = evsel__intval(sample->evsel, sample, kvm_exit_reason(e_machine));
+ key->key = perf_sample__intval(sample, kvm_exit_reason(e_machine));
}
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from trace-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-trace.c | 55 +++++++++++++++++++-------------------
1 file changed, 28 insertions(+), 27 deletions(-)
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index e8aabd7a72fe..eca578212def 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -536,12 +536,12 @@ static struct evsel *perf_evsel__raw_syscall_newtp(const char *direction, void *
return NULL;
}
-#define perf_evsel__sc_tp_uint(evsel, name, sample) \
- ({ struct syscall_tp *fields = __evsel__syscall_tp(evsel); \
+#define perf_evsel__sc_tp_uint(name, sample) \
+ ({ struct syscall_tp *fields = __evsel__syscall_tp(sample->evsel); \
fields->name.integer(&fields->name, sample); })
-#define perf_evsel__sc_tp_ptr(evsel, name, sample) \
- ({ struct syscall_tp *fields = __evsel__syscall_tp(evsel); \
+#define perf_evsel__sc_tp_ptr(name, sample) \
+ ({ struct syscall_tp *fields = __evsel__syscall_tp(sample->evsel); \
fields->name.pointer(&fields->name, sample); })
size_t strarray__scnprintf_suffix(struct strarray *sa, char *bf, size_t size, const char *intfmt, bool show_suffix, int val)
@@ -2749,8 +2749,8 @@ static int trace__printf_interrupted_entry(struct trace *trace)
return printed;
}
-static int trace__fprintf_sample(struct trace *trace, struct evsel *evsel,
- struct perf_sample *sample, struct thread *thread)
+static int trace__fprintf_sample(struct trace *trace, struct perf_sample *sample,
+ struct thread *thread)
{
int printed = 0;
@@ -2758,7 +2758,7 @@ static int trace__fprintf_sample(struct trace *trace, struct evsel *evsel,
double ts = (double)sample->time / NSEC_PER_MSEC;
printed += fprintf(trace->output, "%22s %10.3f %s %d/%d [%d]\n",
- evsel__name(evsel), ts,
+ evsel__name(sample->evsel), ts,
thread__comm_str(thread),
sample->pid, sample->tid, sample->cpu);
}
@@ -2813,7 +2813,7 @@ static int trace__sys_enter(struct trace *trace, struct evsel *evsel,
void *args;
int printed = 0;
struct thread *thread;
- int id = perf_evsel__sc_tp_uint(evsel, id, sample), err = -1;
+ int id = perf_evsel__sc_tp_uint(id, sample), err = -1;
int augmented_args_size = 0, e_machine;
void *augmented_args = NULL;
struct syscall *sc;
@@ -2828,9 +2828,9 @@ static int trace__sys_enter(struct trace *trace, struct evsel *evsel,
if (ttrace == NULL)
goto out_put;
- trace__fprintf_sample(trace, evsel, sample, thread);
+ trace__fprintf_sample(trace, sample, thread);
- args = perf_evsel__sc_tp_ptr(evsel, args, sample);
+ args = perf_evsel__sc_tp_ptr(args, sample);
if (ttrace->entry_str == NULL) {
ttrace->entry_str = malloc(trace__entry_str_size);
@@ -2888,12 +2888,11 @@ static int trace__sys_enter(struct trace *trace, struct evsel *evsel,
return err;
}
-static int trace__fprintf_sys_enter(struct trace *trace, struct evsel *evsel,
- struct perf_sample *sample)
+static int trace__fprintf_sys_enter(struct trace *trace, struct perf_sample *sample)
{
struct thread_trace *ttrace;
struct thread *thread;
- int id = perf_evsel__sc_tp_uint(evsel, id, sample), err = -1;
+ int id = perf_evsel__sc_tp_uint(id, sample), err = -1;
struct syscall *sc;
char msg[1024];
void *args, *augmented_args = NULL;
@@ -2903,7 +2902,7 @@ static int trace__fprintf_sys_enter(struct trace *trace, struct evsel *evsel,
thread = machine__findnew_thread(trace->host, sample->pid, sample->tid);
e_machine = thread__e_machine(thread, trace->host, /*e_flags=*/NULL);
- sc = trace__syscall_info(trace, evsel, e_machine, id);
+ sc = trace__syscall_info(trace, sample->evsel, e_machine, id);
if (sc == NULL)
goto out_put;
ttrace = thread__trace(thread, trace);
@@ -2914,7 +2913,7 @@ static int trace__fprintf_sys_enter(struct trace *trace, struct evsel *evsel,
if (ttrace == NULL)
goto out_put;
- args = perf_evsel__sc_tp_ptr(evsel, args, sample);
+ args = perf_evsel__sc_tp_ptr(args, sample);
augmented_args = syscall__augmented_args(sc, sample, &augmented_args_size, trace->raw_augmented_syscalls_args_size);
printed += syscall__scnprintf_args(sc, msg, sizeof(msg), args, augmented_args, augmented_args_size, trace, thread);
fprintf(trace->output, "%.*s", (int)printed, msg);
@@ -2924,10 +2923,11 @@ static int trace__fprintf_sys_enter(struct trace *trace, struct evsel *evsel,
return err;
}
-static int trace__resolve_callchain(struct trace *trace, struct evsel *evsel,
+static int trace__resolve_callchain(struct trace *trace,
struct perf_sample *sample,
struct callchain_cursor *cursor)
{
+ struct evsel *evsel = sample->evsel;
struct addr_location al;
int max_stack = evsel->core.attr.sample_max_stack ?
evsel->core.attr.sample_max_stack :
@@ -2962,7 +2962,7 @@ static int trace__sys_exit(struct trace *trace, struct evsel *evsel,
u64 duration = 0;
bool duration_calculated = false;
struct thread *thread;
- int id = perf_evsel__sc_tp_uint(evsel, id, sample), err = -1, callchain_ret = 0, printed = 0;
+ int id = perf_evsel__sc_tp_uint(id, sample), err = -1, callchain_ret = 0, printed = 0;
int alignment = trace->args_alignment, e_machine;
struct syscall *sc;
struct thread_trace *ttrace;
@@ -2976,9 +2976,9 @@ static int trace__sys_exit(struct trace *trace, struct evsel *evsel,
if (ttrace == NULL)
goto out_put;
- trace__fprintf_sample(trace, evsel, sample, thread);
+ trace__fprintf_sample(trace, sample, thread);
- ret = perf_evsel__sc_tp_uint(evsel, ret, sample);
+ ret = perf_evsel__sc_tp_uint(ret, sample);
if (trace->summary)
thread__update_stats(thread, ttrace, id, sample, ret, trace);
@@ -3000,7 +3000,7 @@ static int trace__sys_exit(struct trace *trace, struct evsel *evsel,
if (sample->callchain) {
struct callchain_cursor *cursor = get_tls_callchain_cursor();
- callchain_ret = trace__resolve_callchain(trace, evsel, sample, cursor);
+ callchain_ret = trace__resolve_callchain(trace, sample, cursor);
if (callchain_ret == 0) {
if (cursor->nr < trace->min_stack)
goto out;
@@ -3217,9 +3217,10 @@ static void bpf_output__fprintf(struct trace *trace,
++trace->nr_events_printed;
}
-static size_t trace__fprintf_tp_fields(struct trace *trace, struct evsel *evsel, struct perf_sample *sample,
+static size_t trace__fprintf_tp_fields(struct trace *trace, struct perf_sample *sample,
struct thread *thread, void *augmented_args, int augmented_args_size)
{
+ struct evsel *evsel = sample->evsel;
char bf[2048];
size_t size = sizeof(bf);
const struct tep_event *tp_format = evsel__tp_format(evsel);
@@ -3302,7 +3303,7 @@ static int trace__event_handler(struct trace *trace, struct evsel *evsel,
if (sample->callchain) {
struct callchain_cursor *cursor = get_tls_callchain_cursor();
- callchain_ret = trace__resolve_callchain(trace, evsel, sample, cursor);
+ callchain_ret = trace__resolve_callchain(trace, sample, cursor);
if (callchain_ret == 0) {
if (cursor->nr < trace->min_stack)
goto out;
@@ -3323,7 +3324,7 @@ static int trace__event_handler(struct trace *trace, struct evsel *evsel,
trace__fprintf_comm_tid(trace, thread, trace->output);
if (evsel == trace->syscalls.events.bpf_output) {
- int id = perf_evsel__sc_tp_uint(evsel, id, sample);
+ int id = perf_evsel__sc_tp_uint(id, sample);
int e_machine = thread
? thread__e_machine(thread, trace->host, /*e_flags=*/NULL)
: EM_HOST;
@@ -3331,7 +3332,7 @@ static int trace__event_handler(struct trace *trace, struct evsel *evsel,
if (sc) {
fprintf(trace->output, "%s(", sc->name);
- trace__fprintf_sys_enter(trace, evsel, sample);
+ trace__fprintf_sys_enter(trace, sample);
fputc(')', trace->output);
goto newline;
}
@@ -3351,13 +3352,13 @@ static int trace__event_handler(struct trace *trace, struct evsel *evsel,
const struct tep_event *tp_format = evsel__tp_format(evsel);
if (tp_format && (strncmp(tp_format->name, "sys_enter_", 10) ||
- trace__fprintf_sys_enter(trace, evsel, sample))) {
+ trace__fprintf_sys_enter(trace, sample))) {
if (trace->libtraceevent_print) {
event_format__fprintf(tp_format, sample->cpu,
sample->raw_data, sample->raw_size,
trace->output);
} else {
- trace__fprintf_tp_fields(trace, evsel, sample, thread, NULL, 0);
+ trace__fprintf_tp_fields(trace, sample, thread, NULL, 0);
}
}
}
@@ -3416,7 +3417,7 @@ static int trace__pgfault(struct trace *trace,
if (sample->callchain) {
struct callchain_cursor *cursor = get_tls_callchain_cursor();
- callchain_ret = trace__resolve_callchain(trace, evsel, sample, cursor);
+ callchain_ret = trace__resolve_callchain(trace, sample, cursor);
if (callchain_ret == 0) {
if (cursor->nr < trace->min_stack)
goto out_put;
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from callchain-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-c2c.c | 2 +-
tools/perf/builtin-inject.c | 4 ++--
tools/perf/builtin-kmem.c | 6 ++---
tools/perf/builtin-kwork.c | 5 ++--
tools/perf/builtin-lock.c | 24 +++++++++----------
tools/perf/builtin-sched.c | 5 ++--
tools/perf/builtin-script.c | 6 ++---
tools/perf/builtin-trace.c | 2 +-
tools/perf/util/build-id.c | 2 +-
tools/perf/util/callchain.c | 8 +++----
tools/perf/util/callchain.h | 6 ++---
tools/perf/util/db-export.c | 8 +++----
tools/perf/util/hist.c | 2 +-
tools/perf/util/machine.c | 14 +++++------
tools/perf/util/machine.h | 3 ---
.../util/scripting-engines/trace-event-perl.c | 4 ++--
.../scripting-engines/trace-event-python.c | 4 ++--
17 files changed, 48 insertions(+), 57 deletions(-)
diff --git a/tools/perf/builtin-c2c.c b/tools/perf/builtin-c2c.c
index 7593e5908fcc..2fa3d7ec8a09 100644
--- a/tools/perf/builtin-c2c.c
+++ b/tools/perf/builtin-c2c.c
@@ -339,7 +339,7 @@ static int process_sample_event(const struct perf_tool *tool __maybe_unused,
cursor = get_tls_callchain_cursor();
ret = sample__resolve_callchain(sample, cursor, NULL,
- evsel, &al, sysctl_perf_event_max_stack);
+ &al, sysctl_perf_event_max_stack);
if (ret)
goto out;
diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
index 3dac8bcaf6a1..3d48c8140ad9 100644
--- a/tools/perf/builtin-inject.c
+++ b/tools/perf/builtin-inject.c
@@ -418,7 +418,7 @@ static int perf_event__convert_sample_callchain(const struct perf_tool *tool,
goto out;
/* this will parse DWARF using stack and register data */
- ret = thread__resolve_callchain(thread, cursor, evsel, sample,
+ ret = thread__resolve_callchain(thread, cursor, sample,
/*parent=*/NULL, /*root_al=*/NULL,
PERF_MAX_STACK_DEPTH);
thread__put(thread);
@@ -1019,7 +1019,7 @@ static int perf_event__inject_buildid(const struct perf_tool *tool, union perf_e
/*sample_in_dso=*/true);
}
- sample__for_each_callchain_node(thread, sample->evsel, sample, PERF_MAX_STACK_DEPTH,
+ sample__for_each_callchain_node(thread, sample, PERF_MAX_STACK_DEPTH,
/*symbols=*/false, mark_dso_hit_callback, &args);
thread__put(thread);
repipe:
diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index d7233a8b43be..2cdc56bc2616 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -394,7 +394,7 @@ static int build_alloc_func_list(void)
* Find first non-memory allocation function from callchain.
* The allocation functions are in the 'alloc_func_list'.
*/
-static u64 find_callsite(struct evsel *evsel, struct perf_sample *sample)
+static u64 find_callsite(struct perf_sample *sample)
{
struct addr_location al;
struct machine *machine = &kmem_session->machines.host;
@@ -414,7 +414,7 @@ static u64 find_callsite(struct evsel *evsel, struct perf_sample *sample)
if (cursor == NULL)
goto out;
- sample__resolve_callchain(sample, cursor, NULL, evsel, &al, 16);
+ sample__resolve_callchain(sample, cursor, /*parent=*/NULL, &al, 16);
callchain_cursor_commit(cursor);
while (true) {
@@ -838,7 +838,7 @@ static int evsel__process_page_alloc_event(struct evsel *evsel, struct perf_samp
if (parse_gfp_flags(evsel, sample, gfp_flags) < 0)
return -1;
- callsite = find_callsite(evsel, sample);
+ callsite = find_callsite(sample);
/*
* This is to find the current page (with correct gfp flags and
diff --git a/tools/perf/builtin-kwork.c b/tools/perf/builtin-kwork.c
index ab4519287f47..4b312afed830 100644
--- a/tools/perf/builtin-kwork.c
+++ b/tools/perf/builtin-kwork.c
@@ -688,7 +688,6 @@ static int latency_entry_event(struct perf_kwork *kwork,
static void timehist_save_callchain(struct perf_kwork *kwork,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct symbol *sym;
@@ -708,7 +707,7 @@ static void timehist_save_callchain(struct perf_kwork *kwork,
cursor = get_tls_callchain_cursor();
- if (thread__resolve_callchain(thread, cursor, evsel, sample,
+ if (thread__resolve_callchain(thread, cursor, sample,
NULL, NULL, kwork->max_stack + 2) != 0) {
pr_debug("Failed to resolve callchain, skipping\n");
goto out_put;
@@ -838,7 +837,7 @@ static int timehist_entry_event(struct perf_kwork *kwork,
return ret;
if (work != NULL)
- timehist_save_callchain(kwork, sample, evsel, machine);
+ timehist_save_callchain(kwork, sample, machine);
return 0;
}
diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c
index cbf3a39c7837..e2b585e528ac 100644
--- a/tools/perf/builtin-lock.c
+++ b/tools/perf/builtin-lock.c
@@ -551,13 +551,13 @@ static int get_key_by_aggr_mode_simple(u64 *key, u64 addr, u32 tid)
return 0;
}
-static u64 callchain_id(struct evsel *evsel, struct perf_sample *sample);
+static u64 callchain_id(struct perf_sample *sample);
-static int get_key_by_aggr_mode(u64 *key, u64 addr, struct evsel *evsel,
+static int get_key_by_aggr_mode(u64 *key, u64 addr,
struct perf_sample *sample)
{
if (aggr_mode == LOCK_AGGR_CALLER) {
- *key = callchain_id(evsel, sample);
+ *key = callchain_id(sample);
return 0;
}
return get_key_by_aggr_mode_simple(key, addr, sample->tid);
@@ -841,7 +841,7 @@ static int get_symbol_name_offset(struct map *map, struct symbol *sym, u64 ip,
else
return strlcpy(buf, sym->name, size);
}
-static int lock_contention_caller(struct evsel *evsel, struct perf_sample *sample,
+static int lock_contention_caller(struct perf_sample *sample,
char *buf, int size)
{
struct thread *thread;
@@ -862,7 +862,7 @@ static int lock_contention_caller(struct evsel *evsel, struct perf_sample *sampl
cursor = get_tls_callchain_cursor();
/* use caller function name from the callchain */
- ret = thread__resolve_callchain(thread, cursor, evsel, sample,
+ ret = thread__resolve_callchain(thread, cursor, sample,
NULL, NULL, max_stack_depth);
if (ret != 0) {
thread__put(thread);
@@ -896,7 +896,7 @@ static int lock_contention_caller(struct evsel *evsel, struct perf_sample *sampl
return -1;
}
-static u64 callchain_id(struct evsel *evsel, struct perf_sample *sample)
+static u64 callchain_id(struct perf_sample *sample)
{
struct callchain_cursor *cursor;
struct machine *machine = &session->machines.host;
@@ -911,7 +911,7 @@ static u64 callchain_id(struct evsel *evsel, struct perf_sample *sample)
cursor = get_tls_callchain_cursor();
/* use caller function name from the callchain */
- ret = thread__resolve_callchain(thread, cursor, evsel, sample,
+ ret = thread__resolve_callchain(thread, cursor, sample,
NULL, NULL, max_stack_depth);
thread__put(thread);
@@ -963,7 +963,7 @@ static u64 *get_callstack(struct perf_sample *sample, int max_stack)
return callstack;
}
-static int report_lock_contention_begin_event(struct evsel *evsel,
+static int report_lock_contention_begin_event(struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
struct lock_stat *ls;
@@ -978,7 +978,7 @@ static int report_lock_contention_begin_event(struct evsel *evsel,
struct map *kmap;
struct symbol *sym;
- ret = get_key_by_aggr_mode(&key, addr, evsel, sample);
+ ret = get_key_by_aggr_mode(&key, addr, sample);
if (ret < 0)
return ret;
@@ -1025,7 +1025,7 @@ static int report_lock_contention_begin_event(struct evsel *evsel,
break;
case LOCK_AGGR_CALLER:
name = buf;
- if (lock_contention_caller(evsel, sample, buf, sizeof(buf)) < 0)
+ if (lock_contention_caller(sample, buf, sizeof(buf)) < 0)
name = "Unknown";
break;
case LOCK_AGGR_CGROUP:
@@ -1127,7 +1127,7 @@ static int report_lock_contention_begin_event(struct evsel *evsel,
return 0;
}
-static int report_lock_contention_end_event(struct evsel *evsel,
+static int report_lock_contention_end_event(struct evsel *evsel __maybe_unused,
struct perf_sample *sample)
{
struct lock_stat *ls;
@@ -1138,7 +1138,7 @@ static int report_lock_contention_end_event(struct evsel *evsel,
u64 key;
int ret;
- ret = get_key_by_aggr_mode(&key, addr, evsel, sample);
+ ret = get_key_by_aggr_mode(&key, addr, sample);
if (ret < 0)
return ret;
diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
index 01228f630121..2a971081918b 100644
--- a/tools/perf/builtin-sched.c
+++ b/tools/perf/builtin-sched.c
@@ -2325,7 +2325,6 @@ static bool is_idle_sample(struct perf_sample *sample,
static void save_task_callchain(struct perf_sched *sched,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct callchain_cursor *cursor;
@@ -2345,7 +2344,7 @@ static void save_task_callchain(struct perf_sched *sched,
cursor = get_tls_callchain_cursor();
- if (thread__resolve_callchain(thread, cursor, evsel, sample,
+ if (thread__resolve_callchain(thread, cursor, sample,
NULL, NULL, sched->max_stack + 2) != 0) {
if (verbose > 0)
pr_err("Failed to resolve callchain. Skipping\n");
@@ -2519,7 +2518,7 @@ static struct thread *timehist_get_thread(struct perf_sched *sched,
sample->tid);
}
- save_task_callchain(sched, sample, evsel, machine);
+ save_task_callchain(sched, sample, machine);
if (sched->idle_hist) {
struct thread *idle;
struct idle_thread_runtime *itr;
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index d4e8f49a8751..b94a199ffcc9 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -1684,7 +1684,7 @@ static int perf_sample__fprintf_bts(struct perf_sample *sample,
if (symbol_conf.use_callchain && sample->callchain) {
cursor = get_tls_callchain_cursor();
- if (thread__resolve_callchain(al->thread, cursor, evsel,
+ if (thread__resolve_callchain(al->thread, cursor,
sample, NULL, NULL,
scripting_max_stack))
cursor = NULL;
@@ -2507,7 +2507,7 @@ static void process_event(struct perf_script *script,
if (symbol_conf.use_callchain && sample->callchain) {
cursor = get_tls_callchain_cursor();
- if (thread__resolve_callchain(al->thread, cursor, evsel,
+ if (thread__resolve_callchain(al->thread, cursor,
sample, NULL, NULL,
scripting_max_stack))
cursor = NULL;
@@ -2792,7 +2792,7 @@ static int process_deferred_sample_event(const struct perf_tool *tool,
if (symbol_conf.use_callchain && sample->callchain) {
cursor = get_tls_callchain_cursor();
- if (thread__resolve_callchain(al.thread, cursor, evsel,
+ if (thread__resolve_callchain(al.thread, cursor,
sample, NULL, NULL,
scripting_max_stack)) {
pr_info("cannot resolve deferred callchains\n");
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index eca578212def..4bc30ad8fb07 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -2938,7 +2938,7 @@ static int trace__resolve_callchain(struct trace *trace,
if (machine__resolve(trace->host, &al, sample) < 0)
goto out;
- err = thread__resolve_callchain(al.thread, cursor, evsel, sample, NULL, NULL, max_stack);
+ err = thread__resolve_callchain(al.thread, cursor, sample, NULL, NULL, max_stack);
out:
addr_location__exit(&al);
return err;
diff --git a/tools/perf/util/build-id.c b/tools/perf/util/build-id.c
index 55b72235f891..af4d874f1381 100644
--- a/tools/perf/util/build-id.c
+++ b/tools/perf/util/build-id.c
@@ -73,7 +73,7 @@ int build_id__mark_dso_hit(const struct perf_tool *tool __maybe_unused,
addr_location__exit(&al);
- sample__for_each_callchain_node(thread, sample->evsel, sample, PERF_MAX_STACK_DEPTH,
+ sample__for_each_callchain_node(thread, sample, PERF_MAX_STACK_DEPTH,
/*symbols=*/false, mark_dso_hit_callback, /*data=*/NULL);
diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c
index f031cbbeeba8..5c2282051e39 100644
--- a/tools/perf/util/callchain.c
+++ b/tools/perf/util/callchain.c
@@ -1170,7 +1170,7 @@ int callchain_cursor_append(struct callchain_cursor *cursor,
int sample__resolve_callchain(struct perf_sample *sample,
struct callchain_cursor *cursor, struct symbol **parent,
- struct evsel *evsel, struct addr_location *al,
+ struct addr_location *al,
int max_stack)
{
if (sample->callchain == NULL && !symbol_conf.show_branchflag_count)
@@ -1178,7 +1178,7 @@ int sample__resolve_callchain(struct perf_sample *sample,
if (symbol_conf.use_callchain || symbol_conf.cumulate_callchain ||
perf_hpp_list.parent || symbol_conf.show_branchflag_count) {
- return thread__resolve_callchain(al->thread, cursor, evsel, sample,
+ return thread__resolve_callchain(al->thread, cursor, sample,
parent, al, max_stack);
}
return 0;
@@ -1853,7 +1853,7 @@ s64 callchain_avg_cycles(struct callchain_node *cnode)
return cycles;
}
-int sample__for_each_callchain_node(struct thread *thread, struct evsel *evsel,
+int sample__for_each_callchain_node(struct thread *thread,
struct perf_sample *sample, int max_stack,
bool symbols, callchain_iter_fn cb, void *data)
{
@@ -1864,7 +1864,7 @@ int sample__for_each_callchain_node(struct thread *thread, struct evsel *evsel,
return -ENOMEM;
/* Fill in the callchain. */
- ret = __thread__resolve_callchain(thread, cursor, evsel, sample,
+ ret = __thread__resolve_callchain(thread, cursor, sample,
/*parent=*/NULL, /*root_al=*/NULL,
max_stack, symbols);
if (ret)
diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h
index b7702d65ad60..0eb5d7a1a41d 100644
--- a/tools/perf/util/callchain.h
+++ b/tools/perf/util/callchain.h
@@ -8,11 +8,9 @@
#include "branch.h"
struct addr_location;
-struct evsel;
struct hist_entry;
struct hists;
struct ip_callchain;
-struct map;
struct perf_sample;
struct record_opts;
struct thread;
@@ -245,7 +243,7 @@ int record_opts__parse_callchain(struct record_opts *record,
int sample__resolve_callchain(struct perf_sample *sample,
struct callchain_cursor *cursor, struct symbol **parent,
- struct evsel *evsel, struct addr_location *al,
+ struct addr_location *al,
int max_stack);
int hist_entry__append_callchain(struct hist_entry *he, struct perf_sample *sample);
int fill_callchain_info(struct addr_location *al, struct callchain_cursor_node *node,
@@ -306,7 +304,7 @@ s64 callchain_avg_cycles(struct callchain_node *cnode);
typedef int (*callchain_iter_fn)(struct callchain_cursor_node *node, void *data);
-int sample__for_each_callchain_node(struct thread *thread, struct evsel *evsel,
+int sample__for_each_callchain_node(struct thread *thread,
struct perf_sample *sample, int max_stack,
bool symbols, callchain_iter_fn cb, void *data);
diff --git a/tools/perf/util/db-export.c b/tools/perf/util/db-export.c
index cc2bb1af4243..d991e3168aed 100644
--- a/tools/perf/util/db-export.c
+++ b/tools/perf/util/db-export.c
@@ -208,8 +208,7 @@ static int db_ids_from_al(struct db_export *dbe, struct addr_location *al,
static struct call_path *call_path_from_sample(struct db_export *dbe,
struct machine *machine,
struct thread *thread,
- struct perf_sample *sample,
- struct evsel *evsel)
+ struct perf_sample *sample)
{
u64 kernel_start = machine__kernel_start(machine);
struct call_path *current = &dbe->cpr->call_path;
@@ -227,7 +226,7 @@ static struct call_path *call_path_from_sample(struct db_export *dbe,
*/
callchain_param.order = ORDER_CALLER;
cursor = get_tls_callchain_cursor();
- err = thread__resolve_callchain(thread, cursor, evsel,
+ err = thread__resolve_callchain(thread, cursor,
sample, NULL, NULL, PERF_MAX_STACK_DEPTH);
if (err) {
callchain_param.order = saved_order;
@@ -390,8 +389,7 @@ int db_export__sample(struct db_export *dbe, union perf_event *event,
if (dbe->cpr) {
struct call_path *cp = call_path_from_sample(dbe, machine,
- thread, sample,
- evsel);
+ thread, sample);
if (cp) {
db_export__call_path(dbe, cp);
es.call_path_id = cp->db_id;
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index 747fdc455c80..f641cf321ace 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -1342,7 +1342,7 @@ int hist_entry_iter__add(struct hist_entry_iter *iter, struct addr_location *al,
alm = map__get(al->map);
err = sample__resolve_callchain(iter->sample, get_tls_callchain_cursor(), &iter->parent,
- iter->evsel, al, max_stack_depth);
+ al, max_stack_depth);
if (err) {
map__put(alm);
return err;
diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c
index e76f8c86e62a..c2e0a99efe97 100644
--- a/tools/perf/util/machine.c
+++ b/tools/perf/util/machine.c
@@ -2778,13 +2778,13 @@ static u64 get_leaf_frame_caller(struct perf_sample *sample,
static int thread__resolve_callchain_sample(struct thread *thread,
struct callchain_cursor *cursor,
- struct evsel *evsel,
struct perf_sample *sample,
struct symbol **parent,
struct addr_location *root_al,
int max_stack,
bool symbols)
{
+ struct evsel *evsel = sample->evsel;
struct branch_stack *branch = sample->branch_stack;
struct branch_entry *entries = perf_sample__branch_entries(sample);
struct ip_callchain *chain = sample->callchain;
@@ -2986,10 +2986,11 @@ static int unwind_entry(struct unwind_entry *entry, void *arg)
static int thread__resolve_callchain_unwind(struct thread *thread,
struct callchain_cursor *cursor,
- struct evsel *evsel,
struct perf_sample *sample,
int max_stack, bool symbols)
{
+ struct evsel *evsel = sample->evsel;
+
/* Can we do dwarf post unwind? */
if (!((evsel->core.attr.sample_type & PERF_SAMPLE_REGS_USER) &&
(evsel->core.attr.sample_type & PERF_SAMPLE_STACK_USER)))
@@ -3009,7 +3010,6 @@ static int thread__resolve_callchain_unwind(struct thread *thread,
int __thread__resolve_callchain(struct thread *thread,
struct callchain_cursor *cursor,
- struct evsel *evsel,
struct perf_sample *sample,
struct symbol **parent,
struct addr_location *root_al,
@@ -3025,22 +3025,22 @@ int __thread__resolve_callchain(struct thread *thread,
if (callchain_param.order == ORDER_CALLEE) {
ret = thread__resolve_callchain_sample(thread, cursor,
- evsel, sample,
+ sample,
parent, root_al,
max_stack, symbols);
if (ret)
return ret;
ret = thread__resolve_callchain_unwind(thread, cursor,
- evsel, sample,
+ sample,
max_stack, symbols);
} else {
ret = thread__resolve_callchain_unwind(thread, cursor,
- evsel, sample,
+ sample,
max_stack, symbols);
if (ret)
return ret;
ret = thread__resolve_callchain_sample(thread, cursor,
- evsel, sample,
+ sample,
parent, root_al,
max_stack, symbols);
}
diff --git a/tools/perf/util/machine.h b/tools/perf/util/machine.h
index 22a42c5825fa..048b24e9bd38 100644
--- a/tools/perf/util/machine.h
+++ b/tools/perf/util/machine.h
@@ -187,7 +187,6 @@ struct callchain_cursor;
int __thread__resolve_callchain(struct thread *thread,
struct callchain_cursor *cursor,
- struct evsel *evsel,
struct perf_sample *sample,
struct symbol **parent,
struct addr_location *root_al,
@@ -196,7 +195,6 @@ int __thread__resolve_callchain(struct thread *thread,
static inline int thread__resolve_callchain(struct thread *thread,
struct callchain_cursor *cursor,
- struct evsel *evsel,
struct perf_sample *sample,
struct symbol **parent,
struct addr_location *root_al,
@@ -204,7 +202,6 @@ static inline int thread__resolve_callchain(struct thread *thread,
{
return __thread__resolve_callchain(thread,
cursor,
- evsel,
sample,
parent,
root_al,
diff --git a/tools/perf/util/scripting-engines/trace-event-perl.c b/tools/perf/util/scripting-engines/trace-event-perl.c
index e261a57b87d4..af0d514b2397 100644
--- a/tools/perf/util/scripting-engines/trace-event-perl.c
+++ b/tools/perf/util/scripting-engines/trace-event-perl.c
@@ -257,7 +257,7 @@ static void define_event_symbols(struct tep_event *event,
}
static SV *perl_process_callchain(struct perf_sample *sample,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct addr_location *al)
{
struct callchain_cursor *cursor;
@@ -272,7 +272,7 @@ static SV *perl_process_callchain(struct perf_sample *sample,
cursor = get_tls_callchain_cursor();
- if (thread__resolve_callchain(al->thread, cursor, evsel,
+ if (thread__resolve_callchain(al->thread, cursor,
sample, NULL, NULL, scripting_max_stack) != 0) {
pr_err("Failed to resolve callchain. Skipping\n");
goto exit;
diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c
index 5a30caaec73e..1537122167d5 100644
--- a/tools/perf/util/scripting-engines/trace-event-python.c
+++ b/tools/perf/util/scripting-engines/trace-event-python.c
@@ -390,7 +390,7 @@ static unsigned long get_offset(struct symbol *sym, struct addr_location *al)
}
static PyObject *python_process_callchain(struct perf_sample *sample,
- struct evsel *evsel,
+ struct evsel *evsel __maybe_unused,
struct addr_location *al)
{
PyObject *pylist;
@@ -404,7 +404,7 @@ static PyObject *python_process_callchain(struct perf_sample *sample,
goto exit;
cursor = get_tls_callchain_cursor();
- if (thread__resolve_callchain(al->thread, cursor, evsel,
+ if (thread__resolve_callchain(al->thread, cursor,
sample, NULL, NULL,
scripting_max_stack) != 0) {
pr_err("Failed to resolve callchain. Skipping\n");
--
2.54.0.746.g67dd491aae-googThe evsel is within the sample and so only the sample needs to be
passed. Remove the parameter and fix call site.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-lock.c | 65 ++++++++++++++++-----------------------
1 file changed, 26 insertions(+), 39 deletions(-)
diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c
index e2b585e528ac..b6533483ba64 100644
--- a/tools/perf/builtin-lock.c
+++ b/tools/perf/builtin-lock.c
@@ -473,28 +473,22 @@ static struct lock_stat *pop_from_result(void)
struct trace_lock_handler {
/* it's used on CONFIG_LOCKDEP */
- int (*acquire_event)(struct evsel *evsel,
- struct perf_sample *sample);
+ int (*acquire_event)(struct perf_sample *sample);
/* it's used on CONFIG_LOCKDEP && CONFIG_LOCK_STAT */
- int (*acquired_event)(struct evsel *evsel,
- struct perf_sample *sample);
+ int (*acquired_event)(struct perf_sample *sample);
/* it's used on CONFIG_LOCKDEP && CONFIG_LOCK_STAT */
- int (*contended_event)(struct evsel *evsel,
- struct perf_sample *sample);
+ int (*contended_event)(struct perf_sample *sample);
/* it's used on CONFIG_LOCKDEP */
- int (*release_event)(struct evsel *evsel,
- struct perf_sample *sample);
+ int (*release_event)(struct perf_sample *sample);
/* it's used when CONFIG_LOCKDEP is off */
- int (*contention_begin_event)(struct evsel *evsel,
- struct perf_sample *sample);
+ int (*contention_begin_event)(struct perf_sample *sample);
/* it's used when CONFIG_LOCKDEP is off */
- int (*contention_end_event)(struct evsel *evsel,
- struct perf_sample *sample);
+ int (*contention_end_event)(struct perf_sample *sample);
};
static struct lock_seq_stat *get_seq(struct thread_stat *ts, u64 addr)
@@ -563,8 +557,7 @@ static int get_key_by_aggr_mode(u64 *key, u64 addr,
return get_key_by_aggr_mode_simple(key, addr, sample->tid);
}
-static int report_lock_acquire_event(struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+static int report_lock_acquire_event(struct perf_sample *sample)
{
struct lock_stat *ls;
struct thread_stat *ts;
@@ -638,8 +631,7 @@ static int report_lock_acquire_event(struct evsel *evsel __maybe_unused,
return 0;
}
-static int report_lock_acquired_event(struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+static int report_lock_acquired_event(struct perf_sample *sample)
{
struct lock_stat *ls;
struct thread_stat *ts;
@@ -704,8 +696,7 @@ static int report_lock_acquired_event(struct evsel *evsel __maybe_unused,
return 0;
}
-static int report_lock_contended_event(struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+static int report_lock_contended_event(struct perf_sample *sample)
{
struct lock_stat *ls;
struct thread_stat *ts;
@@ -762,8 +753,7 @@ static int report_lock_contended_event(struct evsel *evsel __maybe_unused,
return 0;
}
-static int report_lock_release_event(struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+static int report_lock_release_event(struct perf_sample *sample)
{
struct lock_stat *ls;
struct thread_stat *ts;
@@ -963,8 +953,7 @@ static u64 *get_callstack(struct perf_sample *sample, int max_stack)
return callstack;
}
-static int report_lock_contention_begin_event(struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+static int report_lock_contention_begin_event(struct perf_sample *sample)
{
struct lock_stat *ls;
struct thread_stat *ts;
@@ -1127,8 +1116,7 @@ static int report_lock_contention_begin_event(struct evsel *evsel __maybe_unused
return 0;
}
-static int report_lock_contention_end_event(struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+static int report_lock_contention_end_event(struct perf_sample *sample)
{
struct lock_stat *ls;
struct thread_stat *ts;
@@ -1208,45 +1196,45 @@ static struct trace_lock_handler contention_lock_ops = {
static struct trace_lock_handler *trace_handler;
-static int evsel__process_lock_acquire(struct evsel *evsel, struct perf_sample *sample)
+static int evsel__process_lock_acquire(struct perf_sample *sample)
{
if (trace_handler->acquire_event)
- return trace_handler->acquire_event(evsel, sample);
+ return trace_handler->acquire_event(sample);
return 0;
}
-static int evsel__process_lock_acquired(struct evsel *evsel, struct perf_sample *sample)
+static int evsel__process_lock_acquired(struct perf_sample *sample)
{
if (trace_handler->acquired_event)
- return trace_handler->acquired_event(evsel, sample);
+ return trace_handler->acquired_event(sample);
return 0;
}
-static int evsel__process_lock_contended(struct evsel *evsel, struct perf_sample *sample)
+static int evsel__process_lock_contended(struct perf_sample *sample)
{
if (trace_handler->contended_event)
- return trace_handler->contended_event(evsel, sample);
+ return trace_handler->contended_event(sample);
return 0;
}
-static int evsel__process_lock_release(struct evsel *evsel, struct perf_sample *sample)
+static int evsel__process_lock_release(struct perf_sample *sample)
{
if (trace_handler->release_event)
- return trace_handler->release_event(evsel, sample);
+ return trace_handler->release_event(sample);
return 0;
}
-static int evsel__process_contention_begin(struct evsel *evsel, struct perf_sample *sample)
+static int evsel__process_contention_begin(struct perf_sample *sample)
{
if (trace_handler->contention_begin_event)
- return trace_handler->contention_begin_event(evsel, sample);
+ return trace_handler->contention_begin_event(sample);
return 0;
}
-static int evsel__process_contention_end(struct evsel *evsel, struct perf_sample *sample)
+static int evsel__process_contention_end(struct perf_sample *sample)
{
if (trace_handler->contention_end_event)
- return trace_handler->contention_end_event(evsel, sample);
+ return trace_handler->contention_end_event(sample);
return 0;
}
@@ -1424,8 +1412,7 @@ static int process_event_update(const struct perf_tool *tool,
return 0;
}
-typedef int (*tracepoint_handler)(struct evsel *evsel,
- struct perf_sample *sample);
+typedef int (*tracepoint_handler)(struct perf_sample *sample);
static int process_sample_event(const struct perf_tool *tool __maybe_unused,
union perf_event *event,
@@ -1445,7 +1432,7 @@ static int process_sample_event(const struct perf_tool *tool __maybe_unused,
if (evsel->handler != NULL) {
tracepoint_handler f = evsel->handler;
- err = f(evsel, sample);
+ err = f(sample);
}
thread__put(thread);
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from hist-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-annotate.c | 7 +++----
tools/perf/builtin-c2c.c | 2 +-
tools/perf/builtin-report.c | 18 ++++++++----------
tools/perf/builtin-top.c | 6 +++---
tools/perf/util/annotate.c | 19 +++++++++----------
tools/perf/util/annotate.h | 6 ++----
6 files changed, 26 insertions(+), 32 deletions(-)
diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index 58e56f826367..ee1ba2dc35f4 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -176,16 +176,15 @@ static int hist_iter__branch_callback(struct hist_entry_iter *iter,
struct hist_entry *he = iter->he;
struct branch_info *bi;
struct perf_sample *sample = iter->sample;
- struct evsel *evsel = iter->evsel;
int err;
bi = he->branch_info;
- err = addr_map_symbol__inc_samples(&bi->from, sample, evsel);
+ err = addr_map_symbol__inc_samples(&bi->from, sample);
if (err)
goto out;
- err = addr_map_symbol__inc_samples(&bi->to, sample, evsel);
+ err = addr_map_symbol__inc_samples(&bi->to, sample);
out:
return err;
@@ -275,7 +274,7 @@ static int evsel__add_sample(struct evsel *evsel, struct perf_sample *sample,
if (he == NULL)
return -ENOMEM;
- ret = hist_entry__inc_addr_samples(he, sample, evsel, al->addr);
+ ret = hist_entry__inc_addr_samples(he, sample, al->addr);
hists__inc_nr_samples(hists, true);
return ret;
}
diff --git a/tools/perf/builtin-c2c.c b/tools/perf/builtin-c2c.c
index 2fa3d7ec8a09..36f386949923 100644
--- a/tools/perf/builtin-c2c.c
+++ b/tools/perf/builtin-c2c.c
@@ -371,7 +371,7 @@ static int process_sample_event(const struct perf_tool *tool __maybe_unused,
if (perf_c2c__has_annotation(NULL)) {
perf_c2c__evsel_hists_inc_stats(evsel, he, sample);
- addr_map_symbol__inc_samples(mem_info__iaddr(mi), sample, evsel);
+ addr_map_symbol__inc_samples(mem_info__iaddr(mi), sample);
}
ret = hist_entry__append_callchain(he, sample);
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index 9955ce8cce00..cc3ee712fe62 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -171,7 +171,6 @@ static int hist_iter__report_callback(struct hist_entry_iter *iter,
int err = 0;
struct report *rep = arg;
struct hist_entry *he = iter->he;
- struct evsel *evsel = iter->evsel;
struct perf_sample *sample = iter->sample;
struct mem_info *mi;
struct branch_info *bi;
@@ -181,25 +180,25 @@ static int hist_iter__report_callback(struct hist_entry_iter *iter,
if (sort__mode == SORT_MODE__BRANCH) {
bi = he->branch_info;
- err = addr_map_symbol__inc_samples(&bi->from, sample, evsel);
+ err = addr_map_symbol__inc_samples(&bi->from, sample);
if (err)
goto out;
- err = addr_map_symbol__inc_samples(&bi->to, sample, evsel);
+ err = addr_map_symbol__inc_samples(&bi->to, sample);
} else if (rep->mem_mode) {
mi = he->mem_info;
- err = addr_map_symbol__inc_samples(mem_info__daddr(mi), sample, evsel);
+ err = addr_map_symbol__inc_samples(mem_info__daddr(mi), sample);
if (err)
goto out;
- err = hist_entry__inc_addr_samples(he, sample, evsel, al->addr);
+ err = hist_entry__inc_addr_samples(he, sample, al->addr);
} else if (symbol_conf.cumulate_callchain) {
if (single)
- err = hist_entry__inc_addr_samples(he, sample, evsel, al->addr);
+ err = hist_entry__inc_addr_samples(he, sample, al->addr);
} else {
- err = hist_entry__inc_addr_samples(he, sample, evsel, al->addr);
+ err = hist_entry__inc_addr_samples(he, sample, al->addr);
}
out:
@@ -215,7 +214,6 @@ static int hist_iter__branch_callback(struct hist_entry_iter *iter,
struct report *rep = arg;
struct branch_info *bi = he->branch_info;
struct perf_sample *sample = iter->sample;
- struct evsel *evsel = iter->evsel;
int err;
branch_type_count(&rep->brtype_stat, &bi->flags,
@@ -224,11 +222,11 @@ static int hist_iter__branch_callback(struct hist_entry_iter *iter,
if (!ui__has_annotation() && !rep->symbol_ipc)
return 0;
- err = addr_map_symbol__inc_samples(&bi->from, sample, evsel);
+ err = addr_map_symbol__inc_samples(&bi->from, sample);
if (err)
goto out;
- err = addr_map_symbol__inc_samples(&bi->to, sample, evsel);
+ err = addr_map_symbol__inc_samples(&bi->to, sample);
out:
return err;
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index f6eb543de537..b4fc991b4eeb 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -199,7 +199,7 @@ static void ui__warn_map_erange(struct map *map, struct symbol *sym, u64 ip)
static void perf_top__record_precise_ip(struct perf_top *top,
struct hist_entry *he,
struct perf_sample *sample,
- struct evsel *evsel, u64 ip)
+ u64 ip)
EXCLUSIVE_LOCKS_REQUIRED(he->hists->lock)
{
struct annotation *notes;
@@ -216,7 +216,7 @@ static void perf_top__record_precise_ip(struct perf_top *top,
if (!annotation__trylock(notes))
return;
- err = hist_entry__inc_addr_samples(he, sample, evsel, ip);
+ err = hist_entry__inc_addr_samples(he, sample, ip);
annotation__unlock(notes);
@@ -735,7 +735,7 @@ static int hist_iter__top_callback(struct hist_entry_iter *iter,
struct evsel *evsel = iter->evsel;
if (perf_hpp_list.sym && single)
- perf_top__record_precise_ip(top, iter->he, iter->sample, evsel, al->addr);
+ perf_top__record_precise_ip(top, iter->he, iter->sample, al->addr);
hist__account_cycles(iter->sample->branch_stack, al, iter->sample,
!(top->record_opts.branch_stack & PERF_SAMPLE_BRANCH_ANY),
diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c
index e745f3034a0e..470569745abe 100644
--- a/tools/perf/util/annotate.c
+++ b/tools/perf/util/annotate.c
@@ -213,9 +213,10 @@ static int __symbol__account_cycles(struct cyc_hist *ch,
}
static int __symbol__inc_addr_samples(struct map_symbol *ms,
- struct annotated_source *src, struct evsel *evsel, u64 addr,
+ struct annotated_source *src, u64 addr,
struct perf_sample *sample)
{
+ struct evsel *evsel = sample->evsel;
struct symbol *sym = ms->sym;
long hash_key;
u64 offset;
@@ -318,7 +319,7 @@ struct annotated_source *symbol__hists(struct symbol *sym, int nr_hists)
}
static int symbol__inc_addr_samples(struct map_symbol *ms,
- struct evsel *evsel, u64 addr,
+ u64 addr,
struct perf_sample *sample)
{
struct symbol *sym = ms->sym;
@@ -326,8 +327,8 @@ static int symbol__inc_addr_samples(struct map_symbol *ms,
if (sym == NULL)
return 0;
- src = symbol__hists(sym, evsel->evlist->core.nr_entries);
- return src ? __symbol__inc_addr_samples(ms, src, evsel, addr, sample) : 0;
+ src = symbol__hists(sym, sample->evsel->evlist->core.nr_entries);
+ return src ? __symbol__inc_addr_samples(ms, src, addr, sample) : 0;
}
static int symbol__account_br_cntr(struct annotated_branch *branch,
@@ -581,16 +582,14 @@ static int annotation__compute_ipc(struct annotation *notes, size_t size,
return 0;
}
-int addr_map_symbol__inc_samples(struct addr_map_symbol *ams, struct perf_sample *sample,
- struct evsel *evsel)
+int addr_map_symbol__inc_samples(struct addr_map_symbol *ams, struct perf_sample *sample)
{
- return symbol__inc_addr_samples(&ams->ms, evsel, ams->al_addr, sample);
+ return symbol__inc_addr_samples(&ams->ms, ams->al_addr, sample);
}
-int hist_entry__inc_addr_samples(struct hist_entry *he, struct perf_sample *sample,
- struct evsel *evsel, u64 ip)
+int hist_entry__inc_addr_samples(struct hist_entry *he, struct perf_sample *sample, u64 ip)
{
- return symbol__inc_addr_samples(&he->ms, evsel, ip, sample);
+ return symbol__inc_addr_samples(&he->ms, ip, sample);
}
diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h
index 696e36dbf013..1aa6df7d1618 100644
--- a/tools/perf/util/annotate.h
+++ b/tools/perf/util/annotate.h
@@ -422,8 +422,7 @@ static inline struct annotation *symbol__annotation(struct symbol *sym)
return (void *)sym - symbol_conf.priv_size;
}
-int addr_map_symbol__inc_samples(struct addr_map_symbol *ams, struct perf_sample *sample,
- struct evsel *evsel);
+int addr_map_symbol__inc_samples(struct addr_map_symbol *ams, struct perf_sample *sample);
struct annotated_branch *annotation__get_branch(struct annotation *notes);
@@ -433,8 +432,7 @@ int addr_map_symbol__account_cycles(struct addr_map_symbol *ams,
struct evsel *evsel,
u64 br_cntr);
-int hist_entry__inc_addr_samples(struct hist_entry *he, struct perf_sample *sample,
- struct evsel *evsel, u64 addr);
+int hist_entry__inc_addr_samples(struct hist_entry *he, struct perf_sample *sample, u64 addr);
struct annotated_source *symbol__hists(struct symbol *sym, int nr_hists);
void symbol__annotate_zero_histograms(struct symbol *sym);
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from db-export-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/util/db-export.c | 5 ++---
tools/perf/util/db-export.h | 3 +--
tools/perf/util/scripting-engines/trace-event-python.c | 8 ++++----
3 files changed, 7 insertions(+), 9 deletions(-)
diff --git a/tools/perf/util/db-export.c b/tools/perf/util/db-export.c
index d991e3168aed..ba54b1119ee6 100644
--- a/tools/perf/util/db-export.c
+++ b/tools/perf/util/db-export.c
@@ -344,14 +344,13 @@ static int db_export__threads(struct db_export *dbe, struct thread *thread,
}
int db_export__sample(struct db_export *dbe, union perf_event *event,
- struct perf_sample *sample, struct evsel *evsel,
+ struct perf_sample *sample,
struct addr_location *al, struct addr_location *addr_al)
{
struct thread *thread = al->thread;
struct export_sample es = {
.event = event,
.sample = sample,
- .evsel = evsel,
.al = al,
};
struct thread *main_thread;
@@ -364,7 +363,7 @@ int db_export__sample(struct db_export *dbe, union perf_event *event,
if (!machine)
return -1;
- err = db_export__evsel(dbe, evsel);
+ err = db_export__evsel(dbe, sample->evsel);
if (err)
return err;
diff --git a/tools/perf/util/db-export.h b/tools/perf/util/db-export.h
index 23983cb35706..1abbfd398e3a 100644
--- a/tools/perf/util/db-export.h
+++ b/tools/perf/util/db-export.h
@@ -25,7 +25,6 @@ struct call_return;
struct export_sample {
union perf_event *event;
struct perf_sample *sample;
- struct evsel *evsel;
struct addr_location *al;
u64 db_id;
u64 comm_db_id;
@@ -96,7 +95,7 @@ int db_export__symbol(struct db_export *dbe, struct symbol *sym,
int db_export__branch_type(struct db_export *dbe, u32 branch_type,
const char *name);
int db_export__sample(struct db_export *dbe, union perf_event *event,
- struct perf_sample *sample, struct evsel *evsel,
+ struct perf_sample *sample,
struct addr_location *al, struct addr_location *addr_al);
int db_export__branch_types(struct db_export *dbe);
diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c
index 1537122167d5..9e1069ec0578 100644
--- a/tools/perf/util/scripting-engines/trace-event-python.c
+++ b/tools/perf/util/scripting-engines/trace-event-python.c
@@ -1312,7 +1312,7 @@ static void python_export_sample_table(struct db_export *dbe,
t = tuple_new(28);
tuple_set_d64(t, 0, es->db_id);
- tuple_set_d64(t, 1, es->evsel->db_id);
+ tuple_set_d64(t, 1, es->sample->evsel->db_id);
tuple_set_d64(t, 2, maps__machine(thread__maps(es->al->thread))->db_id);
tuple_set_d64(t, 3, thread__db_id(es->al->thread));
tuple_set_d64(t, 4, es->comm_db_id);
@@ -1353,7 +1353,7 @@ static void python_export_synth(struct db_export *dbe, struct export_sample *es)
t = tuple_new(3);
tuple_set_d64(t, 0, es->db_id);
- tuple_set_d64(t, 1, es->evsel->core.attr.config);
+ tuple_set_d64(t, 1, es->sample->evsel->core.attr.config);
tuple_set_bytes(t, 2, es->sample->raw_data, es->sample->raw_size);
call_object(tables->synth_handler, t, "synth_data");
@@ -1368,7 +1368,7 @@ static int python_export_sample(struct db_export *dbe,
python_export_sample_table(dbe, es);
- if (es->evsel->core.attr.type == PERF_TYPE_SYNTH && tables->synth_handler)
+ if (es->sample->evsel->core.attr.type == PERF_TYPE_SYNTH && tables->synth_handler)
python_export_synth(dbe, es);
return 0;
@@ -1517,7 +1517,7 @@ static void python_process_event(union perf_event *event,
/* Reserve for future process_hw/sw/raw APIs */
default:
if (tables->db_export_mode)
- db_export__sample(&tables->dbe, event, sample, evsel, al, addr_al);
+ db_export__sample(&tables->dbe, event, sample, al, addr_al);
else
python_process_general_event(sample, evsel, al, addr_al);
}
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from hist-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-annotate.c | 10 ++++------
tools/perf/builtin-diff.c | 3 +--
tools/perf/builtin-report.c | 6 ++----
tools/perf/builtin-top.c | 9 +++------
tools/perf/tests/hists_cumulate.c | 1 -
tools/perf/tests/hists_filter.c | 1 -
tools/perf/tests/hists_output.c | 1 -
tools/perf/util/hist.c | 24 ++++++++++++------------
tools/perf/util/hist.h | 3 +--
9 files changed, 23 insertions(+), 35 deletions(-)
diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index ee1ba2dc35f4..6f8be9ead43b 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -190,14 +190,12 @@ static int hist_iter__branch_callback(struct hist_entry_iter *iter,
return err;
}
-static int process_branch_callback(struct evsel *evsel,
- struct perf_sample *sample,
+static int process_branch_callback(struct perf_sample *sample,
struct addr_location *al,
struct perf_annotate *ann,
struct machine *machine)
{
struct hist_entry_iter iter = {
- .evsel = evsel,
.sample = sample,
.add_entry_cb = hist_iter__branch_callback,
.hide_unresolved = symbol_conf.hide_unresolved,
@@ -220,8 +218,8 @@ static int process_branch_callback(struct evsel *evsel,
if (a.map != NULL)
dso__set_hit(map__dso(a.map));
- hist__account_cycles(sample->branch_stack, al, sample, false,
- NULL, evsel);
+ hist__account_cycles(sample->branch_stack, al, sample, /*nonany_branch_mode=*/false,
+ /*total_cycles=*/NULL);
ret = hist_entry_iter__add(&iter, &a, PERF_MAX_STACK_DEPTH, ann);
out:
@@ -268,7 +266,7 @@ static int evsel__add_sample(struct evsel *evsel, struct perf_sample *sample,
process_branch_stack(sample->branch_stack, al, sample);
if (ann->has_br_stack && has_annotation(ann))
- return process_branch_callback(evsel, sample, al, ann, machine);
+ return process_branch_callback(sample, al, ann, machine);
he = hists__add_entry(hists, al, NULL, NULL, NULL, NULL, sample, true);
if (he == NULL)
diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c
index 82e9e514922b..b4ff863b304c 100644
--- a/tools/perf/builtin-diff.c
+++ b/tools/perf/builtin-diff.c
@@ -397,7 +397,6 @@ static int diff__process_sample_event(const struct perf_tool *tool,
struct evsel *evsel = sample->evsel;
struct hists *hists = evsel__hists(evsel);
struct hist_entry_iter iter = {
- .evsel = evsel,
.sample = sample,
.ops = &hist_iter_normal,
};
@@ -431,7 +430,7 @@ static int diff__process_sample_event(const struct perf_tool *tool,
}
hist__account_cycles(sample->branch_stack, &al, sample,
- false, NULL, evsel);
+ /*nonany_branch_mode=*/false, /*total_cycles=*/NULL);
break;
case COMPUTE_STREAM:
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index cc3ee712fe62..11c7eff92b95 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -266,10 +266,8 @@ static int process_sample_event(const struct perf_tool *tool,
struct machine *machine)
{
struct report *rep = container_of(tool, struct report, tool);
- struct evsel *evsel = sample->evsel;
struct addr_location al;
struct hist_entry_iter iter = {
- .evsel = evsel,
.sample = sample,
.hide_unresolved = symbol_conf.hide_unresolved,
.add_entry_cb = hist_iter__report_callback,
@@ -281,7 +279,7 @@ static int process_sample_event(const struct perf_tool *tool,
return 0;
}
- if (evswitch__discard(&rep->evswitch, evsel))
+ if (evswitch__discard(&rep->evswitch, sample->evsel))
return 0;
addr_location__init(&al);
@@ -325,7 +323,7 @@ static int process_sample_event(const struct perf_tool *tool,
if (ui__has_annotation() || rep->symbol_ipc || rep->total_cycles_mode) {
hist__account_cycles(sample->branch_stack, &al, sample,
rep->nonany_branch_mode,
- &rep->total_cycles, evsel);
+ &rep->total_cycles);
}
rep->total_samples++;
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index b4fc991b4eeb..6cf73bb0c7af 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -732,20 +732,18 @@ static int hist_iter__top_callback(struct hist_entry_iter *iter,
EXCLUSIVE_LOCKS_REQUIRED(iter->he->hists->lock)
{
struct perf_top *top = arg;
- struct evsel *evsel = iter->evsel;
if (perf_hpp_list.sym && single)
perf_top__record_precise_ip(top, iter->he, iter->sample, al->addr);
hist__account_cycles(iter->sample->branch_stack, al, iter->sample,
!(top->record_opts.branch_stack & PERF_SAMPLE_BRANCH_ANY),
- NULL, evsel);
+ /*total_cycles=*/NULL);
return 0;
}
static void perf_event__process_sample(const struct perf_tool *tool,
const union perf_event *event,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -831,9 +829,8 @@ static void perf_event__process_sample(const struct perf_tool *tool,
}
if (al.sym == NULL || !al.sym->idle) {
- struct hists *hists = evsel__hists(evsel);
+ struct hists *hists = evsel__hists(sample->evsel);
struct hist_entry_iter iter = {
- .evsel = evsel,
.sample = sample,
.add_entry_cb = hist_iter__top_callback,
};
@@ -1211,7 +1208,7 @@ static int deliver_event(struct ordered_events *qe,
}
if (event->header.type == PERF_RECORD_SAMPLE) {
- perf_event__process_sample(&top->tool, event, evsel,
+ perf_event__process_sample(&top->tool, event,
&sample, machine);
} else if (event->header.type == PERF_RECORD_LOST) {
perf_top__process_lost(top, event, evsel);
diff --git a/tools/perf/tests/hists_cumulate.c b/tools/perf/tests/hists_cumulate.c
index 606aa926a8fc..267cbc24691a 100644
--- a/tools/perf/tests/hists_cumulate.c
+++ b/tools/perf/tests/hists_cumulate.c
@@ -87,7 +87,6 @@ static int add_hist_entries(struct hists *hists, struct machine *machine)
addr_location__init(&al);
for (i = 0; i < ARRAY_SIZE(fake_samples); i++) {
struct hist_entry_iter iter = {
- .evsel = evsel,
.sample = &sample,
.hide_unresolved = false,
};
diff --git a/tools/perf/tests/hists_filter.c b/tools/perf/tests/hists_filter.c
index cc6b26e373d1..002e3a4c1ca5 100644
--- a/tools/perf/tests/hists_filter.c
+++ b/tools/perf/tests/hists_filter.c
@@ -63,7 +63,6 @@ static int add_hist_entries(struct evlist *evlist,
evlist__for_each_entry(evlist, evsel) {
for (i = 0; i < ARRAY_SIZE(fake_samples); i++) {
struct hist_entry_iter iter = {
- .evsel = evsel,
.sample = &sample,
.ops = &hist_iter_normal,
.hide_unresolved = false,
diff --git a/tools/perf/tests/hists_output.c b/tools/perf/tests/hists_output.c
index 7818950d786e..fa683fd7b1e5 100644
--- a/tools/perf/tests/hists_output.c
+++ b/tools/perf/tests/hists_output.c
@@ -57,7 +57,6 @@ static int add_hist_entries(struct hists *hists, struct machine *machine)
addr_location__init(&al);
for (i = 0; i < ARRAY_SIZE(fake_samples); i++) {
struct hist_entry_iter iter = {
- .evsel = evsel,
.sample = &sample,
.ops = &hist_iter_normal,
.hide_unresolved = false,
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index f641cf321ace..811d68fa6770 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -935,8 +935,8 @@ iter_add_single_mem_entry(struct hist_entry_iter *iter, struct addr_location *al
{
u64 cost;
struct mem_info *mi = iter->mi;
- struct hists *hists = evsel__hists(iter->evsel);
struct perf_sample *sample = iter->sample;
+ struct hists *hists = evsel__hists(sample->evsel);
struct hist_entry *he;
if (mi == NULL)
@@ -968,7 +968,7 @@ static int
iter_finish_mem_entry(struct hist_entry_iter *iter,
struct addr_location *al __maybe_unused)
{
- struct evsel *evsel = iter->evsel;
+ struct evsel *evsel = iter->sample->evsel;
struct hists *hists = evsel__hists(evsel);
struct hist_entry *he = iter->he;
int err = -EINVAL;
@@ -1036,9 +1036,9 @@ static int
iter_add_next_branch_entry(struct hist_entry_iter *iter, struct addr_location *al)
{
struct branch_info *bi;
- struct evsel *evsel = iter->evsel;
- struct hists *hists = evsel__hists(evsel);
struct perf_sample *sample = iter->sample;
+ struct evsel *evsel = sample->evsel;
+ struct hists *hists = evsel__hists(evsel);
struct hist_entry *he = NULL;
int i = iter->curr;
int err = 0;
@@ -1078,7 +1078,7 @@ static int
iter_finish_branch_entry(struct hist_entry_iter *iter,
struct addr_location *al __maybe_unused)
{
- struct evsel *evsel = iter->evsel;
+ struct evsel *evsel = iter->sample->evsel;
struct hists *hists = evsel__hists(evsel);
for (int i = 0; i < iter->total; i++)
@@ -1103,8 +1103,8 @@ iter_prepare_normal_entry(struct hist_entry_iter *iter __maybe_unused,
static int
iter_add_single_normal_entry(struct hist_entry_iter *iter, struct addr_location *al)
{
- struct evsel *evsel = iter->evsel;
struct perf_sample *sample = iter->sample;
+ struct evsel *evsel = sample->evsel;
struct hist_entry *he;
he = hists__add_entry(evsel__hists(evsel), al, iter->parent, NULL, NULL,
@@ -1121,8 +1121,8 @@ iter_finish_normal_entry(struct hist_entry_iter *iter,
struct addr_location *al __maybe_unused)
{
struct hist_entry *he = iter->he;
- struct evsel *evsel = iter->evsel;
struct perf_sample *sample = iter->sample;
+ struct evsel *evsel = sample->evsel;
if (he == NULL)
return 0;
@@ -1165,9 +1165,9 @@ static int
iter_add_single_cumulative_entry(struct hist_entry_iter *iter,
struct addr_location *al)
{
- struct evsel *evsel = iter->evsel;
- struct hists *hists = evsel__hists(evsel);
struct perf_sample *sample = iter->sample;
+ struct evsel *evsel = sample->evsel;
+ struct hists *hists = evsel__hists(evsel);
struct hist_entry **he_cache = iter->he_cache;
struct hist_entry *he;
int err = 0;
@@ -1224,8 +1224,8 @@ static int
iter_add_next_cumulative_entry(struct hist_entry_iter *iter,
struct addr_location *al)
{
- struct evsel *evsel = iter->evsel;
struct perf_sample *sample = iter->sample;
+ struct evsel *evsel = sample->evsel;
struct hist_entry **he_cache = iter->he_cache;
struct hist_entry *he;
struct hist_entry he_tmp = {
@@ -2826,7 +2826,7 @@ int hists__unlink(struct hists *hists)
void hist__account_cycles(struct branch_stack *bs, struct addr_location *al,
struct perf_sample *sample, bool nonany_branch_mode,
- u64 *total_cycles, struct evsel *evsel)
+ u64 *total_cycles)
{
struct branch_info *bi;
struct branch_entry *entries = perf_sample__branch_entries(sample);
@@ -2850,7 +2850,7 @@ void hist__account_cycles(struct branch_stack *bs, struct addr_location *al,
for (int i = bs->nr - 1; i >= 0; i--) {
addr_map_symbol__account_cycles(&bi[i].from,
nonany_branch_mode ? NULL : prev,
- bi[i].flags.cycles, evsel,
+ bi[i].flags.cycles, sample->evsel,
bi[i].branch_stack_cntr);
prev = &bi[i].to;
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index d97a4efb9250..8fb89d81ef06 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -156,7 +156,6 @@ struct hist_entry_iter {
int total;
int curr;
- struct evsel *evsel;
struct perf_sample *sample;
struct hist_entry *he;
struct symbol *parent;
@@ -799,7 +798,7 @@ unsigned int hists__overhead_width(struct hists *hists);
void hist__account_cycles(struct branch_stack *bs, struct addr_location *al,
struct perf_sample *sample, bool nonany_branch_mode,
- u64 *total_cycles, struct evsel *evsel);
+ u64 *total_cycles);
struct option;
int parse_filter_percentage(const struct option *opt, const char *arg, int unset);
--
2.54.0.746.g67dd491aae-googIn count_lost_samples_events try to avoid searching for the evsel for
the sample, just use the variable within the sample.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-report.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index 11c7eff92b95..973d97af8501 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -789,9 +789,11 @@ static int count_lost_samples_event(const struct perf_tool *tool,
struct machine *machine __maybe_unused)
{
struct report *rep = container_of(tool, struct report, tool);
- struct evsel *evsel;
+ struct evsel *evsel = sample->evsel;
+
+ if (!evsel)
+ evsel = evlist__id2evsel(rep->session->evlist, sample->id);
- evsel = evlist__id2evsel(rep->session->evlist, sample->id);
if (evsel) {
struct hists *hists = evsel__hists(evsel);
u32 count = event->lost_samples.lost;
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from annotate-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-annotate.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index 6f8be9ead43b..719b36d4eed5 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -232,11 +232,11 @@ static bool has_annotation(struct perf_annotate *ann)
return ui__has_annotation() || ann->use_stdio2;
}
-static int evsel__add_sample(struct evsel *evsel, struct perf_sample *sample,
- struct addr_location *al, struct perf_annotate *ann,
- struct machine *machine)
+static int add_sample(struct perf_sample *sample,
+ struct addr_location *al, struct perf_annotate *ann,
+ struct machine *machine)
{
- struct hists *hists = evsel__hists(evsel);
+ struct hists *hists = evsel__hists(sample->evsel);
struct hist_entry *he;
int ret;
@@ -298,7 +298,7 @@ static int process_sample_event(const struct perf_tool *tool,
goto out_put;
if (!al.filtered &&
- evsel__add_sample(sample->evsel, sample, &al, ann, machine)) {
+ add_sample(sample, &al, ann, machine)) {
pr_warning("problem incrementing symbol count, "
"skipping event\n");
ret = -1;
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from inject-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-inject.c | 52 ++++++++++++++++--------------
tools/perf/util/synthetic-events.c | 30 ++++++++---------
tools/perf/util/synthetic-events.h | 2 --
3 files changed, 40 insertions(+), 44 deletions(-)
diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
index 3d48c8140ad9..f0d7b4888598 100644
--- a/tools/perf/builtin-inject.c
+++ b/tools/perf/builtin-inject.c
@@ -147,14 +147,12 @@ struct event_entry {
static int tool__inject_build_id(const struct perf_tool *tool,
struct perf_sample *sample,
struct machine *machine,
- const struct evsel *evsel,
__u16 misc,
const char *filename,
struct dso *dso, u32 flags);
static int tool__inject_mmap2_build_id(const struct perf_tool *tool,
struct perf_sample *sample,
struct machine *machine,
- const struct evsel *evsel,
__u16 misc,
__u32 pid, __u32 tid,
__u64 start, __u64 len, __u64 pgoff,
@@ -357,7 +355,6 @@ perf_inject__cut_auxtrace_sample(struct perf_inject *inject,
typedef int (*inject_handler)(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine);
static int perf_event__repipe_sample(const struct perf_tool *tool,
@@ -370,7 +367,7 @@ static int perf_event__repipe_sample(const struct perf_tool *tool,
if (evsel && evsel->handler) {
inject_handler f = evsel->handler;
- return f(tool, event, sample, evsel, machine);
+ return f(tool, event, sample, machine);
}
build_id__mark_dso_hit(tool, event, sample, machine);
@@ -585,11 +582,12 @@ static int perf_event__repipe_common_mmap(const struct perf_tool *tool,
}
if (dso && !dso__hit(dso)) {
- struct evsel *evsel = evlist__event2evsel(inject->session->evlist, event);
+ if (!sample->evsel)
+ sample->evsel = evlist__event2evsel(inject->session->evlist, event);
- if (evsel) {
+ if (sample->evsel) {
dso__set_hit(dso);
- tool__inject_build_id(tool, sample, machine, evsel,
+ tool__inject_build_id(tool, sample, machine,
/*misc=*/sample->cpumode,
filename, dso, flags);
}
@@ -616,23 +614,26 @@ static int perf_event__repipe_common_mmap(const struct perf_tool *tool,
}
if ((inject->build_id_style == BID_RWS__MMAP2_BUILDID_ALL) &&
!(event->header.misc & PERF_RECORD_MISC_MMAP_BUILD_ID)) {
- struct evsel *evsel = evlist__event2evsel(inject->session->evlist, event);
+ struct evsel *saved_evsel = sample->evsel;
- if (evsel && !dso_sought) {
+ sample->evsel = evlist__event2evsel(inject->session->evlist, event);
+ if (sample->evsel && !dso_sought) {
dso = findnew_dso(pid, tid, filename, dso_id, machine);
dso_sought = true;
}
- if (evsel && dso &&
- !tool__inject_mmap2_build_id(tool, sample, machine, evsel,
+ if (sample->evsel && dso &&
+ !tool__inject_mmap2_build_id(tool, sample, machine,
sample->cpumode | PERF_RECORD_MISC_MMAP_BUILD_ID,
pid, tid, start, len, pgoff,
dso,
prot, flags,
filename)) {
/* Injected mmap2 so no need to repipe. */
+ sample->evsel = saved_evsel;
dso__put(dso);
return 0;
}
+ sample->evsel = saved_evsel;
}
dso__put(dso);
if (inject->build_id_style == BID_RWS__MMAP2_BUILDID_LAZY)
@@ -837,7 +838,6 @@ static bool perf_inject__lookup_known_build_id(struct perf_inject *inject,
static int tool__inject_build_id(const struct perf_tool *tool,
struct perf_sample *sample,
struct machine *machine,
- const struct evsel *evsel,
__u16 misc,
const char *filename,
struct dso *dso, u32 flags)
@@ -861,7 +861,7 @@ static int tool__inject_build_id(const struct perf_tool *tool,
err = perf_event__synthesize_build_id(tool, sample, machine,
perf_event__repipe,
- evsel, misc, dso__bid(dso),
+ misc, dso__bid(dso),
filename);
if (err) {
pr_err("Can't synthesize build_id event for %s\n", filename);
@@ -874,7 +874,6 @@ static int tool__inject_build_id(const struct perf_tool *tool,
static int tool__inject_mmap2_build_id(const struct perf_tool *tool,
struct perf_sample *sample,
struct machine *machine,
- const struct evsel *evsel,
__u16 misc,
__u32 pid, __u32 tid,
__u64 start, __u64 len, __u64 pgoff,
@@ -897,7 +896,6 @@ static int tool__inject_mmap2_build_id(const struct perf_tool *tool,
err = perf_event__synthesize_mmap2_build_id(tool, sample, machine,
perf_event__repipe,
- evsel,
misc, pid, tid,
start, len, pgoff,
dso__bid(dso),
@@ -914,7 +912,7 @@ static int mark_dso_hit(const struct perf_inject *inject,
const struct perf_tool *tool,
struct perf_sample *sample,
struct machine *machine,
- const struct evsel *mmap_evsel,
+ struct evsel *mmap_evsel,
struct map *map, bool sample_in_dso)
{
struct dso *dso;
@@ -942,9 +940,13 @@ static int mark_dso_hit(const struct perf_inject *inject,
dso = map__dso(map);
if (inject->build_id_style == BID_RWS__INJECT_HEADER_LAZY) {
if (dso && !dso__hit(dso)) {
+ /*
+ * The sample is just read for identifiers which we want
+ * to match the for the event of the sample.
+ */
dso__set_hit(dso);
tool__inject_build_id(tool, sample, machine,
- mmap_evsel, misc, dso__long_name(dso), dso,
+ misc, dso__long_name(dso), dso,
map__flags(map));
}
} else if (inject->build_id_style == BID_RWS__MMAP2_BUILDID_LAZY) {
@@ -952,11 +954,13 @@ static int mark_dso_hit(const struct perf_inject *inject,
const struct build_id null_bid = { .size = 0 };
const struct build_id *bid = dso ? dso__bid(dso) : &null_bid;
const char *filename = dso ? dso__long_name(dso) : "";
+ struct evsel *saved_evsel = sample->evsel;
map__set_hit(map);
+ /* Creating a new mmap2 event which has an evsel for the mmap event. */
+ sample->evsel = mmap_evsel;
perf_event__synthesize_mmap2_build_id(tool, sample, machine,
perf_event__repipe,
- mmap_evsel,
misc,
sample->pid, sample->tid,
map__start(map),
@@ -966,6 +970,7 @@ static int mark_dso_hit(const struct perf_inject *inject,
map__prot(map),
map__flags(map),
filename);
+ sample->evsel = saved_evsel;
}
}
return 0;
@@ -976,7 +981,7 @@ struct mark_dso_hit_args {
const struct perf_tool *tool;
struct perf_sample *sample;
struct machine *machine;
- const struct evsel *mmap_evsel;
+ struct evsel *mmap_evsel;
};
static int mark_dso_hit_callback(struct callchain_cursor_node *node, void *data)
@@ -1031,7 +1036,6 @@ static int perf_event__inject_buildid(const struct perf_tool *tool, union perf_e
static int perf_inject__sched_process_exit(const struct perf_tool *tool,
union perf_event *event __maybe_unused,
struct perf_sample *sample,
- struct evsel *evsel __maybe_unused,
struct machine *machine __maybe_unused)
{
struct perf_inject *inject = container_of(tool, struct perf_inject, tool);
@@ -1051,13 +1055,12 @@ static int perf_inject__sched_process_exit(const struct perf_tool *tool,
static int perf_inject__sched_switch(const struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct perf_inject *inject = container_of(tool, struct perf_inject, tool);
struct event_entry *ent;
- perf_inject__sched_process_exit(tool, event, sample, evsel, machine);
+ perf_inject__sched_process_exit(tool, event, sample, machine);
ent = malloc(event->header.size + sizeof(struct event_entry));
if (ent == NULL) {
@@ -1076,13 +1079,13 @@ static int perf_inject__sched_switch(const struct perf_tool *tool,
static int perf_inject__sched_stat(const struct perf_tool *tool,
union perf_event *event __maybe_unused,
struct perf_sample *sample,
- struct evsel *evsel,
struct machine *machine)
{
struct event_entry *ent;
union perf_event *event_sw;
struct perf_sample sample_sw;
struct perf_inject *inject = container_of(tool, struct perf_inject, tool);
+ struct evsel *evsel = sample->evsel;
u32 pid = perf_sample__intval(sample, "pid");
int ret;
@@ -1447,7 +1450,7 @@ static int synthesize_build_id(struct perf_inject *inject, struct dso *dso, pid_
dso__set_hit(dso);
return perf_event__synthesize_build_id(&inject->tool, &synth_sample, machine,
- process_build_id, inject__mmap_evsel(inject),
+ process_build_id,
/*misc=*/synth_sample.cpumode,
dso__bid(dso), dso__long_name(dso));
}
@@ -2009,7 +2012,6 @@ static int evsel__check_stype(struct evsel *evsel, u64 sample_type, const char *
static int drop_sample(const struct perf_tool *tool __maybe_unused,
union perf_event *event __maybe_unused,
struct perf_sample *sample __maybe_unused,
- struct evsel *evsel __maybe_unused,
struct machine *machine __maybe_unused)
{
return 0;
diff --git a/tools/perf/util/synthetic-events.c b/tools/perf/util/synthetic-events.c
index 85bee747f4cd..de812a2befbc 100644
--- a/tools/perf/util/synthetic-events.c
+++ b/tools/perf/util/synthetic-events.c
@@ -2252,13 +2252,15 @@ int perf_event__synthesize_build_id(const struct perf_tool *tool,
struct perf_sample *sample,
struct machine *machine,
perf_event__handler_t process,
- const struct evsel *evsel,
__u16 misc,
const struct build_id *bid,
const char *filename)
{
union perf_event ev;
size_t len;
+ u64 sample_type = sample->evsel ? sample->evsel->core.attr.sample_type : 0;
+ void *array = &ev;
+ int ret;
len = sizeof(ev.build_id) + strlen(filename) + 1;
len = PERF_ALIGN(len, sizeof(u64));
@@ -2275,23 +2277,17 @@ int perf_event__synthesize_build_id(const struct perf_tool *tool,
ev.build_id.header.size = len;
strcpy(ev.build_id.filename, filename);
- if (evsel) {
- void *array = &ev;
- int ret;
-
- array += ev.header.size;
- ret = perf_event__synthesize_id_sample(array, evsel->core.attr.sample_type, sample);
- if (ret < 0)
- return ret;
-
- if (ret & 7) {
- pr_err("Bad id sample size %d\n", ret);
- return -EINVAL;
- }
+ array += ev.header.size;
+ ret = perf_event__synthesize_id_sample(array, sample_type, sample);
+ if (ret < 0)
+ return ret;
- ev.header.size += ret;
+ if (ret & 7) {
+ pr_err("Bad id sample size %d\n", ret);
+ return -EINVAL;
}
+ ev.header.size += ret;
return process(tool, &ev, sample, machine);
}
@@ -2299,7 +2295,6 @@ int perf_event__synthesize_mmap2_build_id(const struct perf_tool *tool,
struct perf_sample *sample,
struct machine *machine,
perf_event__handler_t process,
- const struct evsel *evsel,
__u16 misc,
__u32 pid, __u32 tid,
__u64 start, __u64 len, __u64 pgoff,
@@ -2309,6 +2304,7 @@ int perf_event__synthesize_mmap2_build_id(const struct perf_tool *tool,
{
union perf_event ev;
size_t ev_len;
+ u64 sample_type = sample->evsel ? sample->evsel->core.attr.sample_type : 0;
void *array;
int ret;
@@ -2339,7 +2335,7 @@ int perf_event__synthesize_mmap2_build_id(const struct perf_tool *tool,
array = &ev;
array += ev.header.size;
- ret = perf_event__synthesize_id_sample(array, evsel->core.attr.sample_type, sample);
+ ret = perf_event__synthesize_id_sample(array, sample_type, sample);
if (ret < 0)
return ret;
diff --git a/tools/perf/util/synthetic-events.h b/tools/perf/util/synthetic-events.h
index b0edad0c3100..473a43a78993 100644
--- a/tools/perf/util/synthetic-events.h
+++ b/tools/perf/util/synthetic-events.h
@@ -50,7 +50,6 @@ int perf_event__synthesize_build_id(const struct perf_tool *tool,
struct perf_sample *sample,
struct machine *machine,
perf_event__handler_t process,
- const struct evsel *evsel,
__u16 misc,
const struct build_id *bid,
const char *filename);
@@ -58,7 +57,6 @@ int perf_event__synthesize_mmap2_build_id(const struct perf_tool *tool,
struct perf_sample *sample,
struct machine *machine,
perf_event__handler_t process,
- const struct evsel *evsel,
__u16 misc,
__u32 pid, __u32 tid,
__u64 start, __u64 len, __u64 pgoff,
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from kmem-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-kmem.c | 26 ++++++++++++--------------
1 file changed, 12 insertions(+), 14 deletions(-)
diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 2cdc56bc2616..68a39f56204d 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -171,7 +171,7 @@ static int insert_caller_stat(unsigned long call_site,
return 0;
}
-static int evsel__process_alloc_event(struct evsel *evsel, struct perf_sample *sample)
+static int evsel__process_alloc_event(struct perf_sample *sample)
{
unsigned long ptr = perf_sample__intval(sample, "ptr"),
call_site = perf_sample__intval(sample, "call_site");
@@ -198,7 +198,7 @@ static int evsel__process_alloc_event(struct evsel *evsel, struct perf_sample *s
* If the tracepoint contains the field "node" the tool stats the
* cross allocation.
*/
- if (evsel__field(evsel, "node")) {
+ if (evsel__field(sample->evsel, "node")) {
int node1, node2;
node1 = cpu__get_node((struct perf_cpu){.cpu = sample->cpu});
@@ -243,7 +243,7 @@ static struct alloc_stat *search_alloc_stat(unsigned long ptr,
return NULL;
}
-static int evsel__process_free_event(struct evsel *evsel __maybe_unused, struct perf_sample *sample)
+static int evsel__process_free_event(struct perf_sample *sample)
{
unsigned long ptr = perf_sample__intval(sample, "ptr");
struct alloc_stat *s_alloc, *s_caller;
@@ -751,8 +751,7 @@ static char *compact_gfp_string(unsigned long gfp_flags)
return NULL;
}
-static int parse_gfp_flags(struct evsel *evsel, struct perf_sample *sample,
- unsigned int gfp_flags)
+static int parse_gfp_flags(struct perf_sample *sample, unsigned int gfp_flags)
{
struct tep_record record = {
.cpu = sample->cpu,
@@ -773,7 +772,7 @@ static int parse_gfp_flags(struct evsel *evsel, struct perf_sample *sample,
}
trace_seq_init(&seq);
- tp_format = evsel__tp_format(evsel);
+ tp_format = evsel__tp_format(sample->evsel);
if (tp_format)
tep_print_event(tp_format->tep, &seq, &record, "%s", TEP_PRINT_INFO);
@@ -805,7 +804,7 @@ static int parse_gfp_flags(struct evsel *evsel, struct perf_sample *sample,
return 0;
}
-static int evsel__process_page_alloc_event(struct evsel *evsel, struct perf_sample *sample)
+static int evsel__process_page_alloc_event(struct perf_sample *sample)
{
u64 page;
unsigned int order = perf_sample__intval(sample, "order");
@@ -835,7 +834,7 @@ static int evsel__process_page_alloc_event(struct evsel *evsel, struct perf_samp
return 0;
}
- if (parse_gfp_flags(evsel, sample, gfp_flags) < 0)
+ if (parse_gfp_flags(sample, gfp_flags) < 0)
return -1;
callsite = find_callsite(sample);
@@ -876,8 +875,7 @@ static int evsel__process_page_alloc_event(struct evsel *evsel, struct perf_samp
return 0;
}
-static int evsel__process_page_free_event(struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+static int evsel__process_page_free_event(struct perf_sample *sample)
{
u64 page;
unsigned int order = perf_sample__intval(sample, "order");
@@ -954,8 +952,7 @@ static bool perf_kmem__skip_sample(struct perf_sample *sample)
return false;
}
-typedef int (*tracepoint_handler)(struct evsel *evsel,
- struct perf_sample *sample);
+typedef int (*tracepoint_handler)(struct perf_sample *sample);
static int process_sample_event(const struct perf_tool *tool __maybe_unused,
union perf_event *event,
@@ -973,14 +970,15 @@ static int process_sample_event(const struct perf_tool *tool __maybe_unused,
return -1;
}
- if (perf_kmem__skip_sample(sample))
+ if (perf_kmem__skip_sample(sample)) {
return 0;
+ }
dump_printf(" ... thread: %s:%d\n", thread__comm_str(thread), thread__tid(thread));
if (evsel->handler != NULL) {
tracepoint_handler f = evsel->handler;
- err = f(evsel, sample);
+ err = f(sample);
}
thread__put(thread);
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from kwork-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-kwork.c | 74 +++++++++++++-------------------------
tools/perf/util/kwork.h | 9 +++--
2 files changed, 28 insertions(+), 55 deletions(-)
diff --git a/tools/perf/builtin-kwork.c b/tools/perf/builtin-kwork.c
index 4b312afed830..59a54d11f7fa 100644
--- a/tools/perf/builtin-kwork.c
+++ b/tools/perf/builtin-kwork.c
@@ -448,7 +448,6 @@ static int work_push_atom(struct perf_kwork *kwork,
struct kwork_class *class,
enum kwork_trace_type src_type,
enum kwork_trace_type dst_type,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine,
struct kwork_work **ret_work,
@@ -458,7 +457,7 @@ static int work_push_atom(struct perf_kwork *kwork,
struct kwork_work *work, key;
BUG_ON(class->work_init == NULL);
- class->work_init(kwork, class, &key, src_type, evsel, sample, machine);
+ class->work_init(kwork, class, &key, src_type, sample, machine);
atom = atom_new(kwork, sample);
if (atom == NULL)
@@ -507,7 +506,6 @@ static struct kwork_atom *work_pop_atom(struct perf_kwork *kwork,
struct kwork_class *class,
enum kwork_trace_type src_type,
enum kwork_trace_type dst_type,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine,
struct kwork_work **ret_work)
@@ -516,7 +514,7 @@ static struct kwork_atom *work_pop_atom(struct perf_kwork *kwork,
struct kwork_work *work, key;
BUG_ON(class->work_init == NULL);
- class->work_init(kwork, class, &key, src_type, evsel, sample, machine);
+ class->work_init(kwork, class, &key, src_type, sample, machine);
work = work_findnew(&class->work_root, &key, &kwork->cmp_id);
if (ret_work != NULL)
@@ -599,18 +597,16 @@ static void report_update_exit_event(struct kwork_work *work,
static int report_entry_event(struct perf_kwork *kwork,
struct kwork_class *class,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
return work_push_atom(kwork, class, KWORK_TRACE_ENTRY,
- KWORK_TRACE_MAX, evsel, sample,
+ KWORK_TRACE_MAX, sample,
machine, NULL, true);
}
static int report_exit_event(struct perf_kwork *kwork,
struct kwork_class *class,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -618,7 +614,7 @@ static int report_exit_event(struct perf_kwork *kwork,
struct kwork_work *work = NULL;
atom = work_pop_atom(kwork, class, KWORK_TRACE_EXIT,
- KWORK_TRACE_ENTRY, evsel, sample,
+ KWORK_TRACE_ENTRY, sample,
machine, &work);
if (work == NULL)
return -1;
@@ -654,18 +650,16 @@ static void latency_update_entry_event(struct kwork_work *work,
static int latency_raise_event(struct perf_kwork *kwork,
struct kwork_class *class,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
return work_push_atom(kwork, class, KWORK_TRACE_RAISE,
- KWORK_TRACE_MAX, evsel, sample,
+ KWORK_TRACE_MAX, sample,
machine, NULL, true);
}
static int latency_entry_event(struct perf_kwork *kwork,
struct kwork_class *class,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -673,7 +667,7 @@ static int latency_entry_event(struct perf_kwork *kwork,
struct kwork_work *work = NULL;
atom = work_pop_atom(kwork, class, KWORK_TRACE_ENTRY,
- KWORK_TRACE_RAISE, evsel, sample,
+ KWORK_TRACE_RAISE, sample,
machine, &work);
if (work == NULL)
return -1;
@@ -812,18 +806,16 @@ static void timehist_print_event(struct perf_kwork *kwork,
static int timehist_raise_event(struct perf_kwork *kwork,
struct kwork_class *class,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
return work_push_atom(kwork, class, KWORK_TRACE_RAISE,
- KWORK_TRACE_MAX, evsel, sample,
+ KWORK_TRACE_MAX, sample,
machine, NULL, true);
}
static int timehist_entry_event(struct perf_kwork *kwork,
struct kwork_class *class,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -831,7 +823,7 @@ static int timehist_entry_event(struct perf_kwork *kwork,
struct kwork_work *work = NULL;
ret = work_push_atom(kwork, class, KWORK_TRACE_ENTRY,
- KWORK_TRACE_RAISE, evsel, sample,
+ KWORK_TRACE_RAISE, sample,
machine, &work, true);
if (ret)
return ret;
@@ -844,7 +836,6 @@ static int timehist_entry_event(struct perf_kwork *kwork,
static int timehist_exit_event(struct perf_kwork *kwork,
struct kwork_class *class,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -861,7 +852,7 @@ static int timehist_exit_event(struct perf_kwork *kwork,
}
atom = work_pop_atom(kwork, class, KWORK_TRACE_EXIT,
- KWORK_TRACE_ENTRY, evsel, sample,
+ KWORK_TRACE_ENTRY, sample,
machine, &work);
if (work == NULL) {
ret = -1;
@@ -895,18 +886,16 @@ static void top_update_runtime(struct kwork_work *work,
static int top_entry_event(struct perf_kwork *kwork,
struct kwork_class *class,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
return work_push_atom(kwork, class, KWORK_TRACE_ENTRY,
- KWORK_TRACE_MAX, evsel, sample,
+ KWORK_TRACE_MAX, sample,
machine, NULL, true);
}
static int top_exit_event(struct perf_kwork *kwork,
struct kwork_class *class,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -915,7 +904,7 @@ static int top_exit_event(struct perf_kwork *kwork,
struct kwork_atom *atom;
atom = work_pop_atom(kwork, class, KWORK_TRACE_EXIT,
- KWORK_TRACE_ENTRY, evsel, sample,
+ KWORK_TRACE_ENTRY, sample,
machine, &work);
if (!work)
return -1;
@@ -936,7 +925,6 @@ static int top_exit_event(struct perf_kwork *kwork,
static int top_sched_switch_event(struct perf_kwork *kwork,
struct kwork_class *class,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -944,7 +932,7 @@ static int top_sched_switch_event(struct perf_kwork *kwork,
struct kwork_work *work;
atom = work_pop_atom(kwork, class, KWORK_TRACE_EXIT,
- KWORK_TRACE_ENTRY, evsel, sample,
+ KWORK_TRACE_ENTRY, sample,
machine, &work);
if (!work)
return -1;
@@ -954,12 +942,11 @@ static int top_sched_switch_event(struct perf_kwork *kwork,
atom_del(atom);
}
- return top_entry_event(kwork, class, evsel, sample, machine);
+ return top_entry_event(kwork, class, sample, machine);
}
static struct kwork_class kwork_irq;
static int process_irq_handler_entry_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -967,12 +954,11 @@ static int process_irq_handler_entry_event(const struct perf_tool *tool,
if (kwork->tp_handler->entry_event)
return kwork->tp_handler->entry_event(kwork, &kwork_irq,
- evsel, sample, machine);
+ sample, machine);
return 0;
}
static int process_irq_handler_exit_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -980,7 +966,7 @@ static int process_irq_handler_exit_event(const struct perf_tool *tool,
if (kwork->tp_handler->exit_event)
return kwork->tp_handler->exit_event(kwork, &kwork_irq,
- evsel, sample, machine);
+ sample, machine);
return 0;
}
@@ -1005,7 +991,6 @@ static void irq_work_init(struct perf_kwork *kwork,
struct kwork_class *class,
struct kwork_work *work,
enum kwork_trace_type src_type __maybe_unused,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine __maybe_unused)
{
@@ -1038,7 +1023,6 @@ static struct kwork_class kwork_irq = {
static struct kwork_class kwork_softirq;
static int process_softirq_raise_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1046,13 +1030,12 @@ static int process_softirq_raise_event(const struct perf_tool *tool,
if (kwork->tp_handler->raise_event)
return kwork->tp_handler->raise_event(kwork, &kwork_softirq,
- evsel, sample, machine);
+ sample, machine);
return 0;
}
static int process_softirq_entry_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1060,13 +1043,12 @@ static int process_softirq_entry_event(const struct perf_tool *tool,
if (kwork->tp_handler->entry_event)
return kwork->tp_handler->entry_event(kwork, &kwork_softirq,
- evsel, sample, machine);
+ sample, machine);
return 0;
}
static int process_softirq_exit_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1074,7 +1056,7 @@ static int process_softirq_exit_event(const struct perf_tool *tool,
if (kwork->tp_handler->exit_event)
return kwork->tp_handler->exit_event(kwork, &kwork_softirq,
- evsel, sample, machine);
+ sample, machine);
return 0;
}
@@ -1133,7 +1115,6 @@ static void softirq_work_init(struct perf_kwork *kwork,
struct kwork_class *class,
struct kwork_work *work,
enum kwork_trace_type src_type __maybe_unused,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine __maybe_unused)
{
@@ -1148,7 +1129,7 @@ static void softirq_work_init(struct perf_kwork *kwork,
} else {
num = perf_sample__intval(sample, "vec");
work->id = num;
- work->name = evsel__softirq_name(evsel, num);
+ work->name = evsel__softirq_name(sample->evsel, num);
}
}
@@ -1169,7 +1150,6 @@ static struct kwork_class kwork_softirq = {
static struct kwork_class kwork_workqueue;
static int process_workqueue_activate_work_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1177,13 +1157,12 @@ static int process_workqueue_activate_work_event(const struct perf_tool *tool,
if (kwork->tp_handler->raise_event)
return kwork->tp_handler->raise_event(kwork, &kwork_workqueue,
- evsel, sample, machine);
+ sample, machine);
return 0;
}
static int process_workqueue_execute_start_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1191,13 +1170,12 @@ static int process_workqueue_execute_start_event(const struct perf_tool *tool,
if (kwork->tp_handler->entry_event)
return kwork->tp_handler->entry_event(kwork, &kwork_workqueue,
- evsel, sample, machine);
+ sample, machine);
return 0;
}
static int process_workqueue_execute_end_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1205,7 +1183,7 @@ static int process_workqueue_execute_end_event(const struct perf_tool *tool,
if (kwork->tp_handler->exit_event)
return kwork->tp_handler->exit_event(kwork, &kwork_workqueue,
- evsel, sample, machine);
+ sample, machine);
return 0;
}
@@ -1233,7 +1211,6 @@ static void workqueue_work_init(struct perf_kwork *kwork __maybe_unused,
struct kwork_class *class,
struct kwork_work *work,
enum kwork_trace_type src_type __maybe_unused,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1267,7 +1244,6 @@ static struct kwork_class kwork_workqueue = {
static struct kwork_class kwork_sched;
static int process_sched_switch_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1275,7 +1251,7 @@ static int process_sched_switch_event(const struct perf_tool *tool,
if (kwork->tp_handler->sched_switch_event)
return kwork->tp_handler->sched_switch_event(kwork, &kwork_sched,
- evsel, sample, machine);
+ sample, machine);
return 0;
}
@@ -1300,7 +1276,6 @@ static void sched_work_init(struct perf_kwork *kwork __maybe_unused,
struct kwork_class *class,
struct kwork_work *work,
enum kwork_trace_type src_type,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine __maybe_unused)
{
@@ -1946,7 +1921,6 @@ static int perf_kwork__report(struct perf_kwork *kwork)
}
typedef int (*tracepoint_handler)(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine);
@@ -1961,7 +1935,7 @@ static int perf_kwork__process_tracepoint_sample(const struct perf_tool *tool,
if (evsel->handler != NULL) {
tracepoint_handler f = evsel->handler;
- err = f(tool, evsel, sample, machine);
+ err = f(tool, sample, machine);
}
return err;
diff --git a/tools/perf/util/kwork.h b/tools/perf/util/kwork.h
index db00269b73f2..abf637d44794 100644
--- a/tools/perf/util/kwork.h
+++ b/tools/perf/util/kwork.h
@@ -157,7 +157,6 @@ struct kwork_class {
struct kwork_class *class,
struct kwork_work *work,
enum kwork_trace_type src_type,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine);
@@ -167,19 +166,19 @@ struct kwork_class {
struct trace_kwork_handler {
int (*raise_event)(struct perf_kwork *kwork,
- struct kwork_class *class, struct evsel *evsel,
+ struct kwork_class *class,
struct perf_sample *sample, struct machine *machine);
int (*entry_event)(struct perf_kwork *kwork,
- struct kwork_class *class, struct evsel *evsel,
+ struct kwork_class *class,
struct perf_sample *sample, struct machine *machine);
int (*exit_event)(struct perf_kwork *kwork,
- struct kwork_class *class, struct evsel *evsel,
+ struct kwork_class *class,
struct perf_sample *sample, struct machine *machine);
int (*sched_switch_event)(struct perf_kwork *kwork,
- struct kwork_class *class, struct evsel *evsel,
+ struct kwork_class *class,
struct perf_sample *sample, struct machine *machine);
};
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from sched-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-sched.c | 92 ++++++++++++++------------------------
1 file changed, 34 insertions(+), 58 deletions(-)
diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
index 2a971081918b..d984e58c7dbf 100644
--- a/tools/perf/builtin-sched.c
+++ b/tools/perf/builtin-sched.c
@@ -129,21 +129,20 @@ typedef int (*sort_fn_t)(struct work_atoms *, struct work_atoms *);
struct perf_sched;
struct trace_sched_handler {
- int (*switch_event)(struct perf_sched *sched, struct evsel *evsel,
- struct perf_sample *sample, struct machine *machine);
+ int (*switch_event)(struct perf_sched *sched, struct perf_sample *sample,
+ struct machine *machine);
- int (*runtime_event)(struct perf_sched *sched, struct evsel *evsel,
- struct perf_sample *sample, struct machine *machine);
+ int (*runtime_event)(struct perf_sched *sched, struct perf_sample *sample,
+ struct machine *machine);
- int (*wakeup_event)(struct perf_sched *sched, struct evsel *evsel,
- struct perf_sample *sample, struct machine *machine);
+ int (*wakeup_event)(struct perf_sched *sched, struct perf_sample *sample,
+ struct machine *machine);
/* PERF_RECORD_FORK event, not sched_process_fork tracepoint */
int (*fork_event)(struct perf_sched *sched, union perf_event *event,
struct machine *machine);
int (*migrate_task_event)(struct perf_sched *sched,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine);
};
@@ -826,7 +825,7 @@ static void test_calibrations(struct perf_sched *sched)
static int
replay_wakeup_event(struct perf_sched *sched,
- struct evsel *evsel, struct perf_sample *sample,
+ struct perf_sample *sample,
struct machine *machine __maybe_unused)
{
const char *comm = perf_sample__strval(sample, "comm");
@@ -834,7 +833,7 @@ replay_wakeup_event(struct perf_sched *sched,
struct task_desc *waker, *wakee;
if (verbose > 0) {
- printf("sched_wakeup event %p\n", evsel);
+ printf("sched_wakeup event %p\n", sample->evsel);
printf(" ... pid %d woke up %s/%d\n", sample->tid, comm, pid);
}
@@ -847,7 +846,6 @@ replay_wakeup_event(struct perf_sched *sched,
}
static int replay_switch_event(struct perf_sched *sched,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine __maybe_unused)
{
@@ -861,7 +859,7 @@ static int replay_switch_event(struct perf_sched *sched,
s64 delta;
if (verbose > 0)
- printf("sched_switch event %p\n", evsel);
+ printf("sched_switch event %p\n", sample->evsel);
if (cpu >= MAX_CPUS || cpu < 0)
return 0;
@@ -1134,7 +1132,6 @@ static void free_work_atoms(struct work_atoms *atoms)
}
static int latency_switch_event(struct perf_sched *sched,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1204,7 +1201,6 @@ static int latency_switch_event(struct perf_sched *sched,
}
static int latency_runtime_event(struct perf_sched *sched,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1239,7 +1235,6 @@ static int latency_runtime_event(struct perf_sched *sched,
}
static int latency_wakeup_event(struct perf_sched *sched,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1300,7 +1295,6 @@ static int latency_wakeup_event(struct perf_sched *sched,
}
static int latency_migrate_task_event(struct perf_sched *sched,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1519,20 +1513,18 @@ static void perf_sched__sort_lat(struct perf_sched *sched)
}
static int process_sched_wakeup_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
if (sched->tp_handler->wakeup_event)
- return sched->tp_handler->wakeup_event(sched, evsel, sample, machine);
+ return sched->tp_handler->wakeup_event(sched, sample, machine);
return 0;
}
static int process_sched_wakeup_ignore(const struct perf_tool *tool __maybe_unused,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample __maybe_unused,
struct machine *machine __maybe_unused)
{
@@ -1626,8 +1618,8 @@ static void print_sched_map(struct perf_sched *sched, struct perf_cpu this_cpu,
}
}
-static int map_switch_event(struct perf_sched *sched, struct evsel *evsel __maybe_unused,
- struct perf_sample *sample, struct machine *machine)
+static int map_switch_event(struct perf_sched *sched, struct perf_sample *sample,
+ struct machine *machine)
{
const u32 next_pid = perf_sample__intval(sample, "next_pid");
const u32 prev_pid = perf_sample__intval(sample, "prev_pid");
@@ -1791,7 +1783,6 @@ static int map_switch_event(struct perf_sched *sched, struct evsel *evsel __mayb
}
static int process_sched_switch_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -1810,21 +1801,20 @@ static int process_sched_switch_event(const struct perf_tool *tool,
}
if (sched->tp_handler->switch_event)
- err = sched->tp_handler->switch_event(sched, evsel, sample, machine);
+ err = sched->tp_handler->switch_event(sched, sample, machine);
sched->curr_pid[this_cpu] = next_pid;
return err;
}
static int process_sched_runtime_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
if (sched->tp_handler->runtime_event)
- return sched->tp_handler->runtime_event(sched, evsel, sample, machine);
+ return sched->tp_handler->runtime_event(sched, sample, machine);
return 0;
}
@@ -1847,20 +1837,18 @@ static int perf_sched__process_fork_event(const struct perf_tool *tool,
}
static int process_sched_migrate_task_event(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
if (sched->tp_handler->migrate_task_event)
- return sched->tp_handler->migrate_task_event(sched, evsel, sample, machine);
+ return sched->tp_handler->migrate_task_event(sched, sample, machine);
return 0;
}
typedef int (*tracepoint_handler)(const struct perf_tool *tool,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine);
@@ -1874,7 +1862,7 @@ static int perf_sched__process_tracepoint_sample(const struct perf_tool *tool __
if (evsel->handler != NULL) {
tracepoint_handler f = evsel->handler;
- err = f(tool, evsel, sample, machine);
+ err = f(tool, sample, machine);
}
return err;
@@ -2313,11 +2301,10 @@ static void timehist_update_runtime_stats(struct thread_runtime *r,
r->total_pre_mig_time += r->dt_pre_mig;
}
-static bool is_idle_sample(struct perf_sample *sample,
- struct evsel *evsel)
+static bool is_idle_sample(struct perf_sample *sample)
{
/* pid 0 == swapper == idle task */
- if (evsel__name_is(evsel, "sched:sched_switch"))
+ if (evsel__name_is(sample->evsel, "sched:sched_switch"))
return perf_sample__intval(sample, "prev_pid") == 0;
return sample->pid == 0;
@@ -2499,12 +2486,11 @@ static void save_idle_callchain(struct perf_sched *sched,
static struct thread *timehist_get_thread(struct perf_sched *sched,
struct perf_sample *sample,
- struct machine *machine,
- struct evsel *evsel)
+ struct machine *machine)
{
struct thread *thread;
- if (is_idle_sample(sample, evsel)) {
+ if (is_idle_sample(sample)) {
thread = get_idle_thread(sample->cpu);
if (thread == NULL)
pr_err("Failed to get idle thread for cpu %d.\n", sample->cpu);
@@ -2547,7 +2533,6 @@ static struct thread *timehist_get_thread(struct perf_sched *sched,
static bool timehist_skip_sample(struct perf_sched *sched,
struct thread *thread,
- struct evsel *evsel,
struct perf_sample *sample)
{
bool rc = false;
@@ -2569,7 +2554,7 @@ static bool timehist_skip_sample(struct perf_sched *sched,
tr = thread__get_runtime(thread);
if (tr && tr->prio != -1)
prio = tr->prio;
- else if (evsel__name_is(evsel, "sched:sched_switch"))
+ else if (evsel__name_is(sample->evsel, "sched:sched_switch"))
prio = perf_sample__intval(sample, "prev_prio");
if (prio != -1 && !test_bit(prio, sched->prio_bitmap)) {
@@ -2579,7 +2564,7 @@ static bool timehist_skip_sample(struct perf_sched *sched,
}
if (sched->idle_hist) {
- if (!evsel__name_is(evsel, "sched:sched_switch"))
+ if (!evsel__name_is(sample->evsel, "sched:sched_switch"))
rc = true;
else if (perf_sample__intval(sample, "prev_pid") != 0 &&
perf_sample__intval(sample, "next_pid") != 0)
@@ -2590,7 +2575,6 @@ static bool timehist_skip_sample(struct perf_sched *sched,
}
static void timehist_print_wakeup_event(struct perf_sched *sched,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine,
struct thread *awakened)
@@ -2603,8 +2587,8 @@ static void timehist_print_wakeup_event(struct perf_sched *sched,
return;
/* show wakeup unless both awakee and awaker are filtered */
- if (timehist_skip_sample(sched, thread, evsel, sample) &&
- timehist_skip_sample(sched, awakened, evsel, sample)) {
+ if (timehist_skip_sample(sched, thread, sample) &&
+ timehist_skip_sample(sched, awakened, sample)) {
thread__put(thread);
return;
}
@@ -2628,7 +2612,6 @@ static void timehist_print_wakeup_event(struct perf_sched *sched,
static int timehist_sched_wakeup_ignore(const struct perf_tool *tool __maybe_unused,
union perf_event *event __maybe_unused,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample __maybe_unused,
struct machine *machine __maybe_unused)
{
@@ -2637,7 +2620,6 @@ static int timehist_sched_wakeup_ignore(const struct perf_tool *tool __maybe_unu
static int timehist_sched_wakeup_event(const struct perf_tool *tool,
union perf_event *event __maybe_unused,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -2663,14 +2645,13 @@ static int timehist_sched_wakeup_event(const struct perf_tool *tool,
/* show wakeups if requested */
if (sched->show_wakeups &&
!perf_time__skip_sample(&sched->ptime, sample->time))
- timehist_print_wakeup_event(sched, evsel, sample, machine, thread);
+ timehist_print_wakeup_event(sched, sample, machine, thread);
thread__put(thread);
return 0;
}
static void timehist_print_migration_event(struct perf_sched *sched,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine,
struct thread *migrated)
@@ -2691,8 +2672,8 @@ static void timehist_print_migration_event(struct perf_sched *sched,
if (thread == NULL)
return;
- if (timehist_skip_sample(sched, thread, evsel, sample) &&
- timehist_skip_sample(sched, migrated, evsel, sample)) {
+ if (timehist_skip_sample(sched, thread, sample) &&
+ timehist_skip_sample(sched, migrated, sample)) {
thread__put(thread);
return;
}
@@ -2726,7 +2707,6 @@ static void timehist_print_migration_event(struct perf_sched *sched,
static int timehist_migrate_task_event(const struct perf_tool *tool,
union perf_event *event __maybe_unused,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -2751,8 +2731,7 @@ static int timehist_migrate_task_event(const struct perf_tool *tool,
/* show migrations if requested */
if (sched->show_migrations) {
- timehist_print_migration_event(sched, evsel, sample,
- machine, thread);
+ timehist_print_migration_event(sched, sample, machine, thread);
}
thread__put(thread);
@@ -2784,7 +2763,6 @@ static void timehist_update_task_prio(struct perf_sample *sample,
static int timehist_sched_change_event(const struct perf_tool *tool,
union perf_event *event,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine)
{
@@ -2808,13 +2786,13 @@ static int timehist_sched_change_event(const struct perf_tool *tool,
if (sched->show_prio || sched->prio_str)
timehist_update_task_prio(sample, machine);
- thread = timehist_get_thread(sched, sample, machine, evsel);
+ thread = timehist_get_thread(sched, sample, machine);
if (thread == NULL) {
rc = -1;
goto out;
}
- if (timehist_skip_sample(sched, thread, evsel, sample))
+ if (timehist_skip_sample(sched, thread, sample))
goto out;
tr = thread__get_runtime(thread);
@@ -2823,7 +2801,7 @@ static int timehist_sched_change_event(const struct perf_tool *tool,
goto out;
}
- tprev = evsel__get_time(evsel, sample->cpu);
+ tprev = evsel__get_time(sample->evsel, sample->cpu);
/*
* If start time given:
@@ -2913,7 +2891,7 @@ static int timehist_sched_change_event(const struct perf_tool *tool,
tr->migrated = 0;
}
- evsel__save_time(evsel, sample->time, sample->cpu);
+ evsel__save_time(sample->evsel, sample->time, sample->cpu);
thread__put(thread);
addr_location__exit(&al);
@@ -2922,11 +2900,10 @@ static int timehist_sched_change_event(const struct perf_tool *tool,
static int timehist_sched_switch_event(const struct perf_tool *tool,
union perf_event *event,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine __maybe_unused)
{
- return timehist_sched_change_event(tool, event, evsel, sample, machine);
+ return timehist_sched_change_event(tool, event, sample, machine);
}
static int process_lost(const struct perf_tool *tool __maybe_unused,
@@ -3174,7 +3151,6 @@ static void timehist_print_summary(struct perf_sched *sched,
typedef int (*sched_handler)(const struct perf_tool *tool,
union perf_event *event,
- struct evsel *evsel,
struct perf_sample *sample,
struct machine *machine);
@@ -3196,7 +3172,7 @@ static int perf_timehist__process_sample(const struct perf_tool *tool,
if (evsel->handler != NULL) {
sched_handler f = evsel->handler;
- err = f(tool, event, evsel, sample, machine);
+ err = f(tool, event, sample, machine);
}
return err;
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from timechart-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-timechart.c | 60 +++++++++++++++-------------------
1 file changed, 26 insertions(+), 34 deletions(-)
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c
index 034e0ba3f9cb..1d24d8738519 100644
--- a/tools/perf/builtin-timechart.c
+++ b/tools/perf/builtin-timechart.c
@@ -567,7 +567,6 @@ static const char *cat_backtrace(union perf_event *event,
}
typedef int (*tracepoint_handler)(struct timechart *tchart,
- struct evsel *evsel,
struct perf_sample *sample,
const char *backtrace);
@@ -588,8 +587,8 @@ static int process_sample_event(const struct perf_tool *tool,
if (evsel->handler != NULL) {
tracepoint_handler f = evsel->handler;
- return f(tchart, evsel, sample,
- cat_backtrace(event, sample, machine));
+
+ return f(tchart, sample, cat_backtrace(event, sample, machine));
}
return 0;
@@ -597,7 +596,6 @@ static int process_sample_event(const struct perf_tool *tool,
static int
process_sample_cpu_idle(struct timechart *tchart __maybe_unused,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace __maybe_unused)
{
@@ -613,7 +611,6 @@ process_sample_cpu_idle(struct timechart *tchart __maybe_unused,
static int
process_sample_cpu_frequency(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace __maybe_unused)
{
@@ -626,7 +623,6 @@ process_sample_cpu_frequency(struct timechart *tchart,
static int
process_sample_sched_wakeup(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace)
{
@@ -640,7 +636,6 @@ process_sample_sched_wakeup(struct timechart *tchart,
static int
process_sample_sched_switch(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace)
{
@@ -656,7 +651,6 @@ process_sample_sched_switch(struct timechart *tchart,
#ifdef SUPPORT_OLD_POWER_EVENTS
static int
process_sample_power_start(struct timechart *tchart __maybe_unused,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace __maybe_unused)
{
@@ -669,7 +663,6 @@ process_sample_power_start(struct timechart *tchart __maybe_unused,
static int
process_sample_power_end(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace __maybe_unused)
{
@@ -679,7 +672,6 @@ process_sample_power_end(struct timechart *tchart,
static int
process_sample_power_frequency(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
struct perf_sample *sample,
const char *backtrace __maybe_unused)
{
@@ -849,8 +841,8 @@ static int pid_end_io_sample(struct timechart *tchart, int pid, int type,
static int
process_enter_read(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+ struct perf_sample *sample,
+ const char *backtrace __maybe_unused)
{
long fd = perf_sample__intval(sample, "fd");
return pid_begin_io_sample(tchart, sample->tid, IOTYPE_READ,
@@ -859,8 +851,8 @@ process_enter_read(struct timechart *tchart,
static int
process_exit_read(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+ struct perf_sample *sample,
+ const char *backtrace __maybe_unused)
{
long ret = perf_sample__intval(sample, "ret");
return pid_end_io_sample(tchart, sample->tid, IOTYPE_READ,
@@ -869,8 +861,8 @@ process_exit_read(struct timechart *tchart,
static int
process_enter_write(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+ struct perf_sample *sample,
+ const char *backtrace __maybe_unused)
{
long fd = perf_sample__intval(sample, "fd");
return pid_begin_io_sample(tchart, sample->tid, IOTYPE_WRITE,
@@ -879,8 +871,8 @@ process_enter_write(struct timechart *tchart,
static int
process_exit_write(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+ struct perf_sample *sample,
+ const char *backtrace __maybe_unused)
{
long ret = perf_sample__intval(sample, "ret");
return pid_end_io_sample(tchart, sample->tid, IOTYPE_WRITE,
@@ -889,8 +881,8 @@ process_exit_write(struct timechart *tchart,
static int
process_enter_sync(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+ struct perf_sample *sample,
+ const char *backtrace __maybe_unused)
{
long fd = perf_sample__intval(sample, "fd");
return pid_begin_io_sample(tchart, sample->tid, IOTYPE_SYNC,
@@ -899,8 +891,8 @@ process_enter_sync(struct timechart *tchart,
static int
process_exit_sync(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+ struct perf_sample *sample,
+ const char *backtrace __maybe_unused)
{
long ret = perf_sample__intval(sample, "ret");
return pid_end_io_sample(tchart, sample->tid, IOTYPE_SYNC,
@@ -909,8 +901,8 @@ process_exit_sync(struct timechart *tchart,
static int
process_enter_tx(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+ struct perf_sample *sample,
+ const char *backtrace __maybe_unused)
{
long fd = perf_sample__intval(sample, "fd");
return pid_begin_io_sample(tchart, sample->tid, IOTYPE_TX,
@@ -919,8 +911,8 @@ process_enter_tx(struct timechart *tchart,
static int
process_exit_tx(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+ struct perf_sample *sample,
+ const char *backtrace __maybe_unused)
{
long ret = perf_sample__intval(sample, "ret");
return pid_end_io_sample(tchart, sample->tid, IOTYPE_TX,
@@ -929,8 +921,8 @@ process_exit_tx(struct timechart *tchart,
static int
process_enter_rx(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+ struct perf_sample *sample,
+ const char *backtrace __maybe_unused)
{
long fd = perf_sample__intval(sample, "fd");
return pid_begin_io_sample(tchart, sample->tid, IOTYPE_RX,
@@ -939,8 +931,8 @@ process_enter_rx(struct timechart *tchart,
static int
process_exit_rx(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+ struct perf_sample *sample,
+ const char *backtrace __maybe_unused)
{
long ret = perf_sample__intval(sample, "ret");
return pid_end_io_sample(tchart, sample->tid, IOTYPE_RX,
@@ -949,8 +941,8 @@ process_exit_rx(struct timechart *tchart,
static int
process_enter_poll(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+ struct perf_sample *sample,
+ const char *backtrace __maybe_unused)
{
long fd = perf_sample__intval(sample, "fd");
return pid_begin_io_sample(tchart, sample->tid, IOTYPE_POLL,
@@ -959,8 +951,8 @@ process_enter_poll(struct timechart *tchart,
static int
process_exit_poll(struct timechart *tchart,
- struct evsel *evsel __maybe_unused,
- struct perf_sample *sample)
+ struct perf_sample *sample,
+ const char *backtrace __maybe_unused)
{
long ret = perf_sample__intval(sample, "ret");
return pid_end_io_sample(tchart, sample->tid, IOTYPE_POLL,
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from trace-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-trace.c | 52 ++++++++++++++++++++------------------
1 file changed, 28 insertions(+), 24 deletions(-)
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index 4bc30ad8fb07..d4fafae853d4 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -2608,7 +2608,7 @@ static struct syscall *trace__find_syscall(struct trace *trace, int e_machine, i
return sc;
}
-typedef int (*tracepoint_handler)(struct trace *trace, struct evsel *evsel,
+typedef int (*tracepoint_handler)(struct trace *trace,
union perf_event *event,
struct perf_sample *sample);
@@ -2805,10 +2805,11 @@ static void *syscall__augmented_args(struct syscall *sc, struct perf_sample *sam
return NULL;
}
-static int trace__sys_enter(struct trace *trace, struct evsel *evsel,
+static int trace__sys_enter(struct trace *trace,
union perf_event *event __maybe_unused,
struct perf_sample *sample)
{
+ struct evsel *evsel = sample->evsel;
char *msg;
void *args;
int printed = 0;
@@ -2954,10 +2955,11 @@ static int trace__fprintf_callchain(struct trace *trace, struct perf_sample *sam
return sample__fprintf_callchain(sample, 38, print_opts, get_tls_callchain_cursor(), symbol_conf.bt_stop_list, trace->output);
}
-static int trace__sys_exit(struct trace *trace, struct evsel *evsel,
+static int trace__sys_exit(struct trace *trace,
union perf_event *event __maybe_unused,
struct perf_sample *sample)
{
+ struct evsel *evsel = sample->evsel;
long ret;
u64 duration = 0;
bool duration_calculated = false;
@@ -3094,7 +3096,7 @@ errno_print: {
return err;
}
-static int trace__vfs_getname(struct trace *trace, struct evsel *evsel __maybe_unused,
+static int trace__vfs_getname(struct trace *trace,
union perf_event *event __maybe_unused,
struct perf_sample *sample)
{
@@ -3155,7 +3157,7 @@ static int trace__vfs_getname(struct trace *trace, struct evsel *evsel __maybe_u
return 0;
}
-static int trace__sched_stat_runtime(struct trace *trace, struct evsel *evsel,
+static int trace__sched_stat_runtime(struct trace *trace,
union perf_event *event __maybe_unused,
struct perf_sample *sample)
{
@@ -3177,7 +3179,7 @@ static int trace__sched_stat_runtime(struct trace *trace, struct evsel *evsel,
out_dump:
fprintf(trace->output, "%s: comm=%s,pid=%u,runtime=%" PRIu64 ",vruntime=%" PRIu64 ")\n",
- evsel->name,
+ sample->evsel->name,
perf_sample__strval(sample, "comm"),
(pid_t)perf_sample__intval(sample, "pid"),
runtime,
@@ -3288,10 +3290,11 @@ static size_t trace__fprintf_tp_fields(struct trace *trace, struct perf_sample *
return fprintf(trace->output, "%.*s", (int)printed, bf);
}
-static int trace__event_handler(struct trace *trace, struct evsel *evsel,
+static int trace__event_handler(struct trace *trace,
union perf_event *event __maybe_unused,
struct perf_sample *sample)
{
+ struct evsel *evsel = sample->evsel;
struct thread *thread;
int callchain_ret = 0;
@@ -3400,7 +3403,6 @@ static void print_location(FILE *f, struct perf_sample *sample,
}
static int trace__pgfault(struct trace *trace,
- struct evsel *evsel,
union perf_event *event __maybe_unused,
struct perf_sample *sample)
{
@@ -3429,7 +3431,7 @@ static int trace__pgfault(struct trace *trace,
if (ttrace == NULL)
goto out_put;
- if (evsel->core.attr.config == PERF_COUNT_SW_PAGE_FAULTS_MAJ) {
+ if (sample->evsel->core.attr.config == PERF_COUNT_SW_PAGE_FAULTS_MAJ) {
ttrace->pfmaj++;
trace->pfmaj++;
} else {
@@ -3446,7 +3448,7 @@ static int trace__pgfault(struct trace *trace,
sample->cpu, trace->output);
fprintf(trace->output, "%sfault [",
- evsel->core.attr.config == PERF_COUNT_SW_PAGE_FAULTS_MAJ ?
+ sample->evsel->core.attr.config == PERF_COUNT_SW_PAGE_FAULTS_MAJ ?
"maj" : "min");
print_location(trace->output, sample, &al, false, true);
@@ -3471,7 +3473,8 @@ static int trace__pgfault(struct trace *trace,
if (callchain_ret > 0)
trace__fprintf_callchain(trace, sample);
else if (callchain_ret < 0)
- pr_err("Problem processing %s callchain, skipping...\n", evsel__name(evsel));
+ pr_err("Problem processing %s callchain, skipping...\n",
+ evsel__name(sample->evsel));
++trace->nr_events_printed;
out:
@@ -3483,7 +3486,6 @@ static int trace__pgfault(struct trace *trace,
}
static void trace__set_base_time(struct trace *trace,
- struct evsel *evsel,
struct perf_sample *sample)
{
/*
@@ -3495,7 +3497,7 @@ static void trace__set_base_time(struct trace *trace,
* appears in our event stream (vfs_getname comes to mind).
*/
if (trace->base_time == 0 && !trace->full_time &&
- (evsel->core.attr.sample_type & PERF_SAMPLE_TIME))
+ (sample->evsel->core.attr.sample_type & PERF_SAMPLE_TIME))
trace->base_time = sample->time;
}
@@ -3515,11 +3517,11 @@ static int trace__process_sample(const struct perf_tool *tool,
if (thread && thread__is_filtered(thread))
goto out;
- trace__set_base_time(trace, evsel, sample);
+ trace__set_base_time(trace, sample);
if (handler) {
++trace->nr_events;
- handler(trace, evsel, event, sample);
+ handler(trace, event, sample);
}
out:
thread__put(thread);
@@ -3661,32 +3663,34 @@ static void evlist__free_syscall_tp_fields(struct evlist *evlist)
static void trace__handle_event(struct trace *trace, union perf_event *event, struct perf_sample *sample)
{
const u32 type = event->header.type;
- struct evsel *evsel;
if (type != PERF_RECORD_SAMPLE) {
trace__process_event(trace, trace->host, event, sample);
return;
}
- evsel = evlist__id2evsel(trace->evlist, sample->id);
- if (evsel == NULL) {
+ if (sample->evsel == NULL)
+ sample->evsel = evlist__id2evsel(trace->evlist, sample->id);
+
+ if (sample->evsel == NULL) {
fprintf(trace->output, "Unknown tp ID %" PRIu64 ", skipping...\n", sample->id);
return;
}
- if (evswitch__discard(&trace->evswitch, evsel))
+ if (evswitch__discard(&trace->evswitch, sample->evsel))
return;
- trace__set_base_time(trace, evsel, sample);
+ trace__set_base_time(trace, sample);
- if (evsel->core.attr.type == PERF_TYPE_TRACEPOINT &&
+ if (sample->evsel->core.attr.type == PERF_TYPE_TRACEPOINT &&
sample->raw_data == NULL) {
fprintf(trace->output, "%s sample with no payload for tid: %d, cpu %d, raw_size=%d, skipping...\n",
- evsel__name(evsel), sample->tid,
+ evsel__name(sample->evsel), sample->tid,
sample->cpu, sample->raw_size);
} else {
- tracepoint_handler handler = evsel->handler;
- handler(trace, evsel, event, sample);
+ tracepoint_handler handler = sample->evsel->handler;
+
+ handler(trace, event, sample);
}
if (trace->nr_events_printed >= trace->max_events && trace->max_events != ULONG_MAX)
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from evlist-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-top.c | 4 +++-
tools/perf/tests/mmap-basic.c | 4 +++-
tools/perf/tests/switch-tracking.c | 5 ++++-
3 files changed, 10 insertions(+), 3 deletions(-)
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index 6cf73bb0c7af..c8474f7ac658 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -1164,7 +1164,9 @@ static int deliver_event(struct ordered_events *qe,
goto next_event;
}
- evsel = evlist__id2evsel(session->evlist, sample.id);
+ evsel = sample.evsel;
+ if (!evsel)
+ evsel = evlist__id2evsel(session->evlist, sample.id);
assert(evsel != NULL);
if (event->header.type == PERF_RECORD_SAMPLE) {
diff --git a/tools/perf/tests/mmap-basic.c b/tools/perf/tests/mmap-basic.c
index 3313c236104e..a18d84d858aa 100644
--- a/tools/perf/tests/mmap-basic.c
+++ b/tools/perf/tests/mmap-basic.c
@@ -142,7 +142,9 @@ static int test__basic_mmap(struct test_suite *test __maybe_unused, int subtest
}
err = -1;
- evsel = evlist__id2evsel(evlist, sample.id);
+ evsel = sample.evsel;
+ if (!evsel)
+ evsel = evlist__id2evsel(evlist, sample.id);
perf_sample__exit(&sample);
if (evsel == NULL) {
pr_debug("event with id %" PRIu64
diff --git a/tools/perf/tests/switch-tracking.c b/tools/perf/tests/switch-tracking.c
index 22b0302252db..e32331fee277 100644
--- a/tools/perf/tests/switch-tracking.c
+++ b/tools/perf/tests/switch-tracking.c
@@ -138,7 +138,10 @@ static int process_sample_event(struct evlist *evlist,
goto out;
}
- evsel = evlist__id2evsel(evlist, sample.id);
+ evsel = sample.evsel;
+ if (!evsel)
+ evsel = evlist__id2evsel(evlist, sample.id);
+
if (evsel == switch_tracking->switch_evsel) {
next_tid = perf_sample__intval(&sample, "next_pid");
prev_tid = perf_sample__intval(&sample, "prev_pid");
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from script-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-script.c | 12 ++++--
.../util/scripting-engines/trace-event-perl.c | 21 +++++------
.../scripting-engines/trace-event-python.c | 37 ++++++++-----------
tools/perf/util/trace-event-scripting.c | 5 +--
tools/perf/util/trace-event.h | 3 --
5 files changed, 33 insertions(+), 45 deletions(-)
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index b94a199ffcc9..c0918006e0ab 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -2419,12 +2419,13 @@ static bool show_event(struct perf_sample *sample,
}
static void process_event(struct perf_script *script,
- struct perf_sample *sample, struct evsel *evsel,
+ struct perf_sample *sample,
struct addr_location *al,
struct addr_location *addr_al,
struct machine *machine)
{
struct thread *thread = al->thread;
+ struct evsel *evsel = sample->evsel;
struct perf_event_attr *attr = &evsel->core.attr;
unsigned int type = evsel__output_type(evsel);
struct evsel_script *es = evsel->priv;
@@ -2716,9 +2717,9 @@ static int process_sample_event(const struct perf_tool *tool,
thread__resolve(al.thread, &addr_al, sample);
addr_al_ptr = &addr_al;
}
- scripting_ops->process_event(event, sample, evsel, &al, addr_al_ptr);
+ scripting_ops->process_event(event, sample, &al, addr_al_ptr);
} else {
- process_event(scr, sample, evsel, &al, &addr_al, machine);
+ process_event(scr, sample, &al, &addr_al, machine);
}
out_put:
@@ -2894,9 +2895,12 @@ static int print_event_with_time(const struct perf_tool *tool,
{
struct perf_script *script = container_of(tool, struct perf_script, tool);
struct perf_session *session = script->session;
- struct evsel *evsel = evlist__id2evsel(session->evlist, sample->id);
+ struct evsel *evsel = sample->evsel;
struct thread *thread = NULL;
+ if (!evsel)
+ evsel = evlist__id2evsel(session->evlist, sample->id);
+
if (evsel && !evsel->core.attr.sample_id_all) {
sample->cpu = 0;
sample->time = timestamp;
diff --git a/tools/perf/util/scripting-engines/trace-event-perl.c b/tools/perf/util/scripting-engines/trace-event-perl.c
index af0d514b2397..7a18ea4b7d50 100644
--- a/tools/perf/util/scripting-engines/trace-event-perl.c
+++ b/tools/perf/util/scripting-engines/trace-event-perl.c
@@ -257,7 +257,6 @@ static void define_event_symbols(struct tep_event *event,
}
static SV *perl_process_callchain(struct perf_sample *sample,
- struct evsel *evsel __maybe_unused,
struct addr_location *al)
{
struct callchain_cursor *cursor;
@@ -340,7 +339,6 @@ static SV *perl_process_callchain(struct perf_sample *sample,
}
static void perl_process_tracepoint(struct perf_sample *sample,
- struct evsel *evsel,
struct addr_location *al)
{
struct thread *thread = al->thread;
@@ -355,6 +353,7 @@ static void perl_process_tracepoint(struct perf_sample *sample,
unsigned long long nsecs = sample->time;
const char *comm = thread__comm_str(thread);
DECLARE_BITMAP(events_defined, TRACE_EVENT_TYPE_MAX);
+ struct evsel *evsel = sample->evsel;
bitmap_zero(events_defined, TRACE_EVENT_TYPE_MAX);
dSP;
@@ -389,7 +388,7 @@ static void perl_process_tracepoint(struct perf_sample *sample,
XPUSHs(sv_2mortal(newSVuv(ns)));
XPUSHs(sv_2mortal(newSViv(pid)));
XPUSHs(sv_2mortal(newSVpv(comm, 0)));
- XPUSHs(sv_2mortal(perl_process_callchain(sample, evsel, al)));
+ XPUSHs(sv_2mortal(perl_process_callchain(sample, al)));
/* common fields other than pid can be accessed via xsub fns */
@@ -426,7 +425,7 @@ static void perl_process_tracepoint(struct perf_sample *sample,
XPUSHs(sv_2mortal(newSVuv(nsecs)));
XPUSHs(sv_2mortal(newSViv(pid)));
XPUSHs(sv_2mortal(newSVpv(comm, 0)));
- XPUSHs(sv_2mortal(perl_process_callchain(sample, evsel, al)));
+ XPUSHs(sv_2mortal(perl_process_callchain(sample, al)));
call_pv("main::trace_unhandled", G_SCALAR);
}
SPAGAIN;
@@ -435,9 +434,7 @@ static void perl_process_tracepoint(struct perf_sample *sample,
LEAVE;
}
-static void perl_process_event_generic(union perf_event *event,
- struct perf_sample *sample,
- struct evsel *evsel)
+static void perl_process_event_generic(union perf_event *event, struct perf_sample *sample)
{
dSP;
@@ -448,7 +445,8 @@ static void perl_process_event_generic(union perf_event *event,
SAVETMPS;
PUSHMARK(SP);
XPUSHs(sv_2mortal(newSVpvn((const char *)event, event->header.size)));
- XPUSHs(sv_2mortal(newSVpvn((const char *)&evsel->core.attr, sizeof(evsel->core.attr))));
+ XPUSHs(sv_2mortal(newSVpvn((const char *)&sample->evsel->core.attr,
+ sizeof(sample->evsel->core.attr))));
XPUSHs(sv_2mortal(newSVpvn((const char *)sample, sizeof(*sample))));
XPUSHs(sv_2mortal(newSVpvn((const char *)sample->raw_data, sample->raw_size)));
PUTBACK;
@@ -461,13 +459,12 @@ static void perl_process_event_generic(union perf_event *event,
static void perl_process_event(union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct addr_location *al,
struct addr_location *addr_al)
{
- scripting_context__update(scripting_context, event, sample, evsel, al, addr_al);
- perl_process_tracepoint(sample, evsel, al);
- perl_process_event_generic(event, sample, evsel);
+ scripting_context__update(scripting_context, event, sample, al, addr_al);
+ perl_process_tracepoint(sample, al);
+ perl_process_event_generic(event, sample);
}
static void run_start_sub(void)
diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c
index 9e1069ec0578..8edd2f36e5a9 100644
--- a/tools/perf/util/scripting-engines/trace-event-python.c
+++ b/tools/perf/util/scripting-engines/trace-event-python.c
@@ -390,7 +390,6 @@ static unsigned long get_offset(struct symbol *sym, struct addr_location *al)
}
static PyObject *python_process_callchain(struct perf_sample *sample,
- struct evsel *evsel __maybe_unused,
struct addr_location *al)
{
PyObject *pylist;
@@ -651,11 +650,9 @@ static PyObject *get_sample_value_as_tuple(struct sample_read_value *value,
return t;
}
-static void set_sample_read_in_dict(PyObject *dict_sample,
- struct perf_sample *sample,
- struct evsel *evsel)
+static void set_sample_read_in_dict(PyObject *dict_sample, struct perf_sample *sample)
{
- u64 read_format = evsel->core.attr.read_format;
+ u64 read_format = sample->evsel->core.attr.read_format;
PyObject *values;
unsigned int i;
@@ -741,11 +738,10 @@ static void regs_map(struct regs_dump *regs, uint64_t mask, uint16_t e_machine,
static int set_regs_in_dict(PyObject *dict,
struct perf_sample *sample,
- struct evsel *evsel,
uint16_t e_machine,
uint32_t e_flags)
{
- struct perf_event_attr *attr = &evsel->core.attr;
+ struct perf_event_attr *attr = &sample->evsel->core.attr;
int size = (__sw_hweight64(attr->sample_regs_intr) * MAX_REG_SIZE) + 1;
char *bf = NULL;
@@ -831,7 +827,6 @@ static void python_process_sample_flags(struct perf_sample *sample, PyObject *di
}
static PyObject *get_perf_sample_dict(struct perf_sample *sample,
- struct evsel *evsel,
struct addr_location *al,
struct addr_location *addr_al,
PyObject *callchain)
@@ -839,6 +834,7 @@ static PyObject *get_perf_sample_dict(struct perf_sample *sample,
PyObject *dict, *dict_sample, *brstack, *brstacksym;
uint16_t e_machine = EM_HOST;
uint32_t e_flags = EF_HOST;
+ struct evsel *evsel = sample->evsel;
dict = PyDict_New();
if (!dict)
@@ -871,7 +867,7 @@ static PyObject *get_perf_sample_dict(struct perf_sample *sample,
PyLong_FromUnsignedLongLong(sample->phys_addr));
pydict_set_item_string_decref(dict_sample, "addr",
PyLong_FromUnsignedLongLong(sample->addr));
- set_sample_read_in_dict(dict_sample, sample, evsel);
+ set_sample_read_in_dict(dict_sample, sample);
pydict_set_item_string_decref(dict_sample, "weight",
PyLong_FromUnsignedLongLong(sample->weight));
pydict_set_item_string_decref(dict_sample, "ins_lat",
@@ -928,7 +924,7 @@ static PyObject *get_perf_sample_dict(struct perf_sample *sample,
if (al->thread)
e_machine = thread__e_machine(al->thread, /*machine=*/NULL, &e_flags);
- if (set_regs_in_dict(dict, sample, evsel, e_machine, e_flags))
+ if (set_regs_in_dict(dict, sample, e_machine, e_flags))
Py_FatalError("Failed to setting regs in dict");
return dict;
@@ -936,7 +932,6 @@ static PyObject *get_perf_sample_dict(struct perf_sample *sample,
#ifdef HAVE_LIBTRACEEVENT
static void python_process_tracepoint(struct perf_sample *sample,
- struct evsel *evsel,
struct addr_location *al,
struct addr_location *addr_al)
{
@@ -954,6 +949,7 @@ static void python_process_tracepoint(struct perf_sample *sample,
const char *comm = thread__comm_str(al->thread);
const char *default_handler_name = "trace_unhandled";
DECLARE_BITMAP(events_defined, TRACE_EVENT_TYPE_MAX);
+ struct evsel *evsel = sample->evsel;
bitmap_zero(events_defined, TRACE_EVENT_TYPE_MAX);
@@ -995,7 +991,7 @@ static void python_process_tracepoint(struct perf_sample *sample,
PyTuple_SetItem(t, n++, context);
/* ip unwinding */
- callchain = python_process_callchain(sample, evsel, al);
+ callchain = python_process_callchain(sample, al);
/* Need an additional reference for the perf_sample dict */
Py_INCREF(callchain);
@@ -1051,7 +1047,7 @@ static void python_process_tracepoint(struct perf_sample *sample,
PyTuple_SetItem(t, n++, dict);
if (get_argument_count(handler) == (int) n + 1) {
- all_entries_dict = get_perf_sample_dict(sample, evsel, al, addr_al,
+ all_entries_dict = get_perf_sample_dict(sample, al, addr_al,
callchain);
PyTuple_SetItem(t, n++, all_entries_dict);
} else {
@@ -1070,7 +1066,6 @@ static void python_process_tracepoint(struct perf_sample *sample,
}
#else
static void python_process_tracepoint(struct perf_sample *sample __maybe_unused,
- struct evsel *evsel __maybe_unused,
struct addr_location *al __maybe_unused,
struct addr_location *addr_al __maybe_unused)
{
@@ -1465,7 +1460,6 @@ static int python_process_call_return(struct call_return *cr, u64 *parent_db_id,
}
static void python_process_general_event(struct perf_sample *sample,
- struct evsel *evsel,
struct addr_location *al,
struct addr_location *addr_al)
{
@@ -1488,8 +1482,8 @@ static void python_process_general_event(struct perf_sample *sample,
Py_FatalError("couldn't create Python tuple");
/* ip unwinding */
- callchain = python_process_callchain(sample, evsel, al);
- dict = get_perf_sample_dict(sample, evsel, al, addr_al, callchain);
+ callchain = python_process_callchain(sample, al);
+ dict = get_perf_sample_dict(sample, al, addr_al, callchain);
PyTuple_SetItem(t, n++, dict);
if (_PyTuple_Resize(&t, n) == -1)
@@ -1502,24 +1496,23 @@ static void python_process_general_event(struct perf_sample *sample,
static void python_process_event(union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct addr_location *al,
struct addr_location *addr_al)
{
struct tables *tables = &tables_global;
- scripting_context__update(scripting_context, event, sample, evsel, al, addr_al);
+ scripting_context__update(scripting_context, event, sample, al, addr_al);
- switch (evsel->core.attr.type) {
+ switch (sample->evsel->core.attr.type) {
case PERF_TYPE_TRACEPOINT:
- python_process_tracepoint(sample, evsel, al, addr_al);
+ python_process_tracepoint(sample, al, addr_al);
break;
/* Reserve for future process_hw/sw/raw APIs */
default:
if (tables->db_export_mode)
db_export__sample(&tables->dbe, event, sample, al, addr_al);
else
- python_process_general_event(sample, evsel, al, addr_al);
+ python_process_general_event(sample, al, addr_al);
}
}
diff --git a/tools/perf/util/trace-event-scripting.c b/tools/perf/util/trace-event-scripting.c
index fa850e44cb46..dc584ac316a3 100644
--- a/tools/perf/util/trace-event-scripting.c
+++ b/tools/perf/util/trace-event-scripting.c
@@ -103,12 +103,11 @@ int script_spec__for_each(int (*cb)(struct scripting_ops *ops, const char *spec)
void scripting_context__update(struct scripting_context *c,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct addr_location *al,
struct addr_location *addr_al)
{
#ifdef HAVE_LIBTRACEEVENT
- const struct tep_event *tp_format = evsel__tp_format(evsel);
+ const struct tep_event *tp_format = evsel__tp_format(sample->evsel);
c->pevent = tp_format ? tp_format->tep : NULL;
#else
@@ -117,7 +116,6 @@ void scripting_context__update(struct scripting_context *c,
c->event_data = sample->raw_data;
c->event = event;
c->sample = sample;
- c->evsel = evsel;
c->al = al;
c->addr_al = addr_al;
}
@@ -134,7 +132,6 @@ static int stop_script_unsupported(void)
static void process_event_unsupported(union perf_event *event __maybe_unused,
struct perf_sample *sample __maybe_unused,
- struct evsel *evsel __maybe_unused,
struct addr_location *al __maybe_unused,
struct addr_location *addr_al __maybe_unused)
{
diff --git a/tools/perf/util/trace-event.h b/tools/perf/util/trace-event.h
index 914d9b69ed62..720121c74f1d 100644
--- a/tools/perf/util/trace-event.h
+++ b/tools/perf/util/trace-event.h
@@ -94,7 +94,6 @@ struct scripting_ops {
int (*stop_script) (void);
void (*process_event) (union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct addr_location *al,
struct addr_location *addr_al);
void (*process_switch)(union perf_event *event,
@@ -124,7 +123,6 @@ struct scripting_context {
void *event_data;
union perf_event *event;
struct perf_sample *sample;
- struct evsel *evsel;
struct addr_location *al;
struct addr_location *addr_al;
struct perf_session *session;
@@ -133,7 +131,6 @@ struct scripting_context {
void scripting_context__update(struct scripting_context *scripting_context,
union perf_event *event,
struct perf_sample *sample,
- struct evsel *evsel,
struct addr_location *al,
struct addr_location *addr_al);
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from s390-sample-raw-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/util/s390-sample-raw.c | 31 +++++++++++++++++--------------
1 file changed, 17 insertions(+), 14 deletions(-)
diff --git a/tools/perf/util/s390-sample-raw.c b/tools/perf/util/s390-sample-raw.c
index c6ae0ae8d86a..52bbca5c56c8 100644
--- a/tools/perf/util/s390-sample-raw.c
+++ b/tools/perf/util/s390-sample-raw.c
@@ -222,8 +222,9 @@ static char *get_counter_name(int set, int nr, struct perf_pmu *pmu)
return result;
}
-static void s390_cpumcfdg_dump(struct perf_pmu *pmu, struct perf_sample *sample)
+static void s390_cpumcfdg_dump(struct perf_sample *sample)
{
+ struct perf_pmu *pmu = sample->evsel->pmu;
size_t i, len = sample->raw_size, offset = 0;
unsigned char *buf = sample->raw_data;
const char *color = PERF_COLOR_BLUE;
@@ -284,8 +285,9 @@ static bool s390_pai_all_test(struct perf_sample *sample)
return true;
}
-static void s390_pai_all_dump(struct evsel *evsel, struct perf_sample *sample)
+static void s390_pai_all_dump(struct perf_sample *sample)
{
+ struct evsel *evsel = sample->evsel;
size_t len = sample->raw_size, offset = 0;
unsigned char *p = sample->raw_data;
const char *color = PERF_COLOR_BLUE;
@@ -332,31 +334,32 @@ void evlist__s390_sample_raw(struct evlist *evlist, union perf_event *event,
struct perf_sample *sample)
{
const char *pai_name;
- struct evsel *evsel;
if (event->header.type != PERF_RECORD_SAMPLE)
return;
- evsel = evlist__event2evsel(evlist, event);
- if (!evsel)
- return;
+ if (!sample->evsel) {
+ sample->evsel = evlist__event2evsel(evlist, event);
+ if (!sample->evsel)
+ return;
+ }
/* Check for raw data in sample */
if (!sample->raw_size || !sample->raw_data)
return;
/* Display raw data on screen */
- if (evsel->core.attr.config == PERF_EVENT_CPUM_CF_DIAG) {
- if (!evsel->pmu)
- evsel->pmu = perf_pmus__find("cpum_cf");
+ if (sample->evsel->core.attr.config == PERF_EVENT_CPUM_CF_DIAG) {
+ if (!sample->evsel->pmu)
+ sample->evsel->pmu = perf_pmus__find("cpum_cf");
if (!s390_cpumcfdg_testctr(sample))
pr_err("Invalid counter set data encountered\n");
else
- s390_cpumcfdg_dump(evsel->pmu, sample);
+ s390_cpumcfdg_dump(sample);
return;
}
- switch (evsel->core.attr.config) {
+ switch (sample->evsel->core.attr.config) {
case PERF_EVENT_PAI_NNPA_ALL:
pai_name = "NNPA_ALL";
break;
@@ -370,8 +373,8 @@ void evlist__s390_sample_raw(struct evlist *evlist, union perf_event *event,
if (!s390_pai_all_test(sample)) {
pr_err("Invalid %s raw data encountered\n", pai_name);
} else {
- if (!evsel->pmu)
- evsel->pmu = perf_pmus__find_by_type(evsel->core.attr.type);
- s390_pai_all_dump(evsel, sample);
+ if (!sample->evsel->pmu)
+ sample->evsel->pmu = perf_pmus__find_by_type(sample->evsel->core.attr.type);
+ s390_pai_all_dump(sample);
}
}
--
2.54.0.746.g67dd491aae-googAs struct perf_sample now directly contains its own resolved evsel pointer,
passing the evsel separately is redundant and clutters the interface.
Remove the redundant evsel parameter from evsel-specific handlers and
structures, ensuring the tool always directly accesses the evsel bound to the
sample. This simplifies the API signatures and eliminates the risk of passing
an inconsistent evsel.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/util/evsel.c | 61 +++++++++++++++++++++++++----------------
1 file changed, 37 insertions(+), 24 deletions(-)
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index ee30e15af054..bb48568b8101 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -3003,52 +3003,62 @@ int evsel__open_per_thread(struct evsel *evsel, struct perf_thread_map *threads)
return ret;
}
-static int perf_evsel__parse_id_sample(const struct evsel *evsel,
- const union perf_event *event,
+static int perf_evsel__parse_id_sample(const union perf_event *event,
struct perf_sample *sample)
{
+ const struct evsel *evsel = sample->evsel;
u64 type = evsel->core.attr.sample_type;
const __u64 *array = event->sample.array;
bool swapped = evsel->needs_swap;
union u64_swap u;
-
- array += ((event->header.size -
- sizeof(event->header)) / sizeof(u64)) - 1;
+ int i = ((event->header.size - sizeof(event->header)) / sizeof(u64)) - 1;
if (type & PERF_SAMPLE_IDENTIFIER) {
- sample->id = *array;
- array--;
+ if (i < 0)
+ return -EFAULT;
+
+ sample->id = array[i--];
}
if (type & PERF_SAMPLE_CPU) {
- u.val64 = *array;
+ if (i < 0)
+ return -EFAULT;
+
+ u.val64 = array[i--];
if (swapped) {
/* undo swap of u64, then swap on individual u32s */
u.val64 = bswap_64(u.val64);
u.val32[0] = bswap_32(u.val32[0]);
}
-
sample->cpu = u.val32[0];
- array--;
}
if (type & PERF_SAMPLE_STREAM_ID) {
- sample->stream_id = *array;
- array--;
+ if (i < 0)
+ return -EFAULT;
+
+ sample->stream_id = array[i--];
}
if (type & PERF_SAMPLE_ID) {
- sample->id = *array;
- array--;
+ if (i < 0)
+ return -EFAULT;
+
+ sample->id = array[i--];
}
if (type & PERF_SAMPLE_TIME) {
- sample->time = *array;
- array--;
+ if (i < 0)
+ return -EFAULT;
+
+ sample->time = array[i--];
}
if (type & PERF_SAMPLE_TID) {
- u.val64 = *array;
+ if (i < 0)
+ return -EFAULT;
+
+ u.val64 = array[i--];
if (swapped) {
/* undo swap of u64, then swap on individual u32s */
u.val64 = bswap_64(u.val64);
@@ -3058,7 +3068,6 @@ static int perf_evsel__parse_id_sample(const struct evsel *evsel,
sample->pid = u.val32[0];
sample->tid = u.val32[1];
- array--;
}
return 0;
@@ -3244,15 +3253,18 @@ int evsel__parse_sample(struct evsel *evsel, union perf_event *event,
data->deferred_cookie = event->callchain_deferred.cookie;
- if (evsel->core.attr.sample_id_all)
- perf_evsel__parse_id_sample(evsel, event, data);
-
+ if (evsel->core.attr.sample_id_all) {
+ if (perf_evsel__parse_id_sample(event, data))
+ goto out_efault;
+ }
return 0;
}
if (event->header.type != PERF_RECORD_SAMPLE) {
- if (evsel->core.attr.sample_id_all)
- perf_evsel__parse_id_sample(evsel, event, data);
+ if (evsel->core.attr.sample_id_all) {
+ if (perf_evsel__parse_id_sample(event, data))
+ goto out_efault;
+ }
return 0;
}
@@ -3614,12 +3626,13 @@ int evsel__parse_sample_timestamp(struct evsel *evsel, union perf_event *event,
if (event->header.type != PERF_RECORD_SAMPLE) {
struct perf_sample data = {
+ .evsel = evsel,
.time = -1ULL,
};
if (!evsel->core.attr.sample_id_all)
return -1;
- if (perf_evsel__parse_id_sample(evsel, event, &data))
+ if (perf_evsel__parse_id_sample(event, &data))
return -1;
*timestamp = data.time;
--
2.54.0.746.g67dd491aae-googConstify the static trace_lock_handler structures.
Since the trace lock handler callbacks and definitions do not change at runtime,
declare them const to enforce read-only memory placement and improve safety.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-lock.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c
index b6533483ba64..89a40d385b27 100644
--- a/tools/perf/builtin-lock.c
+++ b/tools/perf/builtin-lock.c
@@ -1179,7 +1179,7 @@ static int report_lock_contention_end_event(struct perf_sample *sample)
/* lock oriented handlers */
/* TODO: handlers for CPU oriented, thread oriented */
-static struct trace_lock_handler report_lock_ops = {
+static const struct trace_lock_handler report_lock_ops = {
.acquire_event = report_lock_acquire_event,
.acquired_event = report_lock_acquired_event,
.contended_event = report_lock_contended_event,
@@ -1188,13 +1188,13 @@ static struct trace_lock_handler report_lock_ops = {
.contention_end_event = report_lock_contention_end_event,
};
-static struct trace_lock_handler contention_lock_ops = {
+static const struct trace_lock_handler contention_lock_ops = {
.contention_begin_event = report_lock_contention_begin_event,
.contention_end_event = report_lock_contention_end_event,
};
-static struct trace_lock_handler *trace_handler;
+static const struct trace_lock_handler *trace_handler;
static int evsel__process_lock_acquire(struct perf_sample *sample)
{
--
2.54.0.746.g67dd491aae-googAvoid a potential segmentation fault if a parsed sample unexpectedly lacks a
valid callchain pointer.
Add a check for a NULL sample->callchain pointer in get_callstack(). If the
callchain is missing, return NULL to safely skip the event and log a debug
warning rather than causing a tool segfault.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-lock.c | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c
index 89a40d385b27..064b3aa4bad7 100644
--- a/tools/perf/builtin-lock.c
+++ b/tools/perf/builtin-lock.c
@@ -938,9 +938,16 @@ static u64 *get_callstack(struct perf_sample *sample, int max_stack)
u64 i;
int c;
+ if (!sample->callchain) {
+ pr_debug("Sample unexpectedly missing callchain\n");
+ return NULL;
+ }
+
callstack = calloc(max_stack, sizeof(*callstack));
- if (callstack == NULL)
+ if (callstack == NULL) {
+ pr_debug("Failed to allocate callstack\n");
return NULL;
+ }
for (i = 0, c = 0; i < sample->callchain->nr && c < max_stack; i++) {
u64 ip = sample->callchain->ips[i];
@@ -1059,7 +1066,7 @@ static int report_lock_contention_begin_event(struct perf_sample *sample)
if (needs_callstack()) {
u64 *callstack = get_callstack(sample, max_stack_depth);
if (callstack == NULL)
- return -ENOMEM;
+ return 0;
if (!match_callstack_filter(machine, callstack, max_stack_depth)) {
free(callstack);
--
2.54.0.746.g67dd491aae-googResolve a major execution-time leak of backtrace strings in the timechart tool.
- Modify cat_backtrace() to return dynamically allocated memory via
open_memstream(), transferring ownership to the caller.
- Free the returned backtrace string inside process_sample_event() immediately
after invoking the tracepoint handler.
- In handlers like pid_put_sample() and sched_wakeup(), make a separate copy of
the backtrace using strdup() if it needs to be persisted in the sample struct,
preventing lifetime issues and double-free vulnerabilities.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-timechart.c | 17 ++++++++++-------
1 file changed, 10 insertions(+), 7 deletions(-)
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c
index 1d24d8738519..d2e15d02304e 100644
--- a/tools/perf/builtin-timechart.c
+++ b/tools/perf/builtin-timechart.c
@@ -299,7 +299,7 @@ static void pid_put_sample(struct timechart *tchart, int pid, int type,
sample->type = type;
sample->next = c->samples;
sample->cpu = cpu;
- sample->backtrace = backtrace;
+ sample->backtrace = backtrace ? strdup(backtrace) : NULL;
c->samples = sample;
if (sample->type == TYPE_RUNNING && end > start && start > 0) {
@@ -433,7 +433,7 @@ static void sched_wakeup(struct timechart *tchart, int cpu, u64 timestamp,
we->time = timestamp;
we->waker = waker;
- we->backtrace = backtrace;
+ we->backtrace = backtrace ? strdup(backtrace) : NULL;
if ((flags & TRACE_FLAG_HARDIRQ) || (flags & TRACE_FLAG_SOFTIRQ))
we->waker = -1;
@@ -489,9 +489,9 @@ static void sched_switch(struct timechart *tchart, int cpu, u64 timestamp,
}
}
-static const char *cat_backtrace(union perf_event *event,
- struct perf_sample *sample,
- struct machine *machine)
+static char *cat_backtrace(union perf_event *event,
+ struct perf_sample *sample,
+ struct machine *machine)
{
struct addr_location al;
unsigned int i;
@@ -577,6 +577,7 @@ static int process_sample_event(const struct perf_tool *tool,
{
struct timechart *tchart = container_of(tool, struct timechart, tool);
struct evsel *evsel = sample->evsel;
+ int ret = 0;
if (evsel->core.attr.sample_type & PERF_SAMPLE_TIME) {
if (!tchart->first_time || tchart->first_time > sample->time)
@@ -587,11 +588,13 @@ static int process_sample_event(const struct perf_tool *tool,
if (evsel->handler != NULL) {
tracepoint_handler f = evsel->handler;
+ char *backtrace = cat_backtrace(event, sample, machine);
- return f(tchart, sample, cat_backtrace(event, sample, machine));
+ ret = f(tchart, sample, backtrace);
+ free(backtrace);
}
- return 0;
+ return ret;
}
static int
--
2.54.0.746.g67dd491aae-googFix memory leaks on the error paths and skipped sample handling paths
in the perf kmem tool.
Ensure that all allocated GFP flags and thread references are properly freed and
released via thread__put() when skipping samples or encountering parsing failures,
preventing long-term memory usage leaks during large trace analyses.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-kmem.c | 18 +++++++++++++-----
1 file changed, 13 insertions(+), 5 deletions(-)
diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 68a39f56204d..daf2272c7337 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -783,17 +783,21 @@ static int parse_gfp_flags(struct perf_sample *sample, unsigned int gfp_flags)
new = realloc(gfps, (nr_gfps + 1) * sizeof(*gfps));
if (new == NULL)
- return -ENOMEM;
+ goto err_out;
gfps = new;
- new += nr_gfps++;
+ new += nr_gfps;
new->flags = gfp_flags;
new->human_readable = strdup(str + 10);
+ if (!new->human_readable)
+ goto err_out;
new->compact_str = compact_gfp_flags(str + 10);
- if (!new->human_readable || !new->compact_str)
- return -ENOMEM;
-
+ if (!new->compact_str) {
+ free(new->human_readable);
+ goto err_out;
+ }
+ nr_gfps++;
qsort(gfps, nr_gfps, sizeof(*gfps), gfpcmp);
}
@@ -802,6 +806,9 @@ static int parse_gfp_flags(struct perf_sample *sample, unsigned int gfp_flags)
trace_seq_destroy(&seq);
return 0;
+err_out:
+ trace_seq_destroy(&seq);
+ return -ENOMEM;
}
static int evsel__process_page_alloc_event(struct perf_sample *sample)
@@ -971,6 +978,7 @@ static int process_sample_event(const struct perf_tool *tool __maybe_unused,
}
if (perf_kmem__skip_sample(sample)) {
+ thread__put(thread);
return 0;
}
--
2.54.0.746.g67dd491aae-googAdd robust boundary checks when synthesizing mmap2 and build_id events to
ensure that filename fields do not overflow the fixed-size stack allocations or the
synthesized event structures.
Verify that the filename fits safely within the allocated boundaries of the
mmap2 event structure, and prevent potential heap/stack overflow corruptions
from excessively long or corrupted kernel filenames.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/util/synthetic-events.c | 19 ++++++++++++++++---
1 file changed, 16 insertions(+), 3 deletions(-)
diff --git a/tools/perf/util/synthetic-events.c b/tools/perf/util/synthetic-events.c
index de812a2befbc..0816f893b916 100644
--- a/tools/perf/util/synthetic-events.c
+++ b/tools/perf/util/synthetic-events.c
@@ -2257,14 +2257,20 @@ int perf_event__synthesize_build_id(const struct perf_tool *tool,
const char *filename)
{
union perf_event ev;
- size_t len;
+ size_t len, filename_len = strlen(filename);
u64 sample_type = sample->evsel ? sample->evsel->core.attr.sample_type : 0;
void *array = &ev;
int ret;
- len = sizeof(ev.build_id) + strlen(filename) + 1;
+ if (filename_len >= PATH_MAX)
+ return -EINVAL;
+
+ len = sizeof(ev.build_id) + filename_len + 1;
len = PERF_ALIGN(len, sizeof(u64));
+ if (len + MAX_ID_HDR_ENTRIES * sizeof(__u64) > sizeof(ev))
+ return -E2BIG;
+
memset(&ev, 0, len);
ev.build_id.size = bid->size;
@@ -2303,14 +2309,21 @@ int perf_event__synthesize_mmap2_build_id(const struct perf_tool *tool,
const char *filename)
{
union perf_event ev;
+ size_t filename_len = strlen(filename);
size_t ev_len;
u64 sample_type = sample->evsel ? sample->evsel->core.attr.sample_type : 0;
void *array;
int ret;
- ev_len = sizeof(ev.mmap2) - sizeof(ev.mmap2.filename) + strlen(filename) + 1;
+ if (filename_len >= sizeof(ev.mmap2.filename))
+ return -EINVAL;
+
+ ev_len = sizeof(ev.mmap2) - sizeof(ev.mmap2.filename) + filename_len + 1;
ev_len = PERF_ALIGN(ev_len, sizeof(u64));
+ if (ev_len + MAX_ID_HDR_ENTRIES * sizeof(__u64) > sizeof(ev))
+ return -E2BIG;
+
memset(&ev, 0, ev_len);
ev.mmap2.header.type = PERF_RECORD_MMAP2;
--
2.54.0.746.g67dd491aae-googSanitize order and migrate_type values from tracepoint payloads before using
them as array indexes.
When processing page_alloc_event and page_free_event, verify that 'order' is less
than MAX_PAGE_ORDER and 'migrate_type' is less than MAX_MIGRATE_TYPES. This
guarantees that indexing into order_stats[MAX_PAGE_ORDER][MAX_MIGRATE_TYPES] remains
strictly within bounds, avoiding out-of-bound heap or static segment accesses.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-kmem.c | 15 +++++++++++++++
1 file changed, 15 insertions(+)
diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index daf2272c7337..33585e353efe 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -826,6 +826,16 @@ static int evsel__process_page_alloc_event(struct perf_sample *sample)
.migrate_type = migrate_type,
};
+ if (order >= MAX_PAGE_ORDER) {
+ pr_debug("Out-of-bounds order %u\n", order);
+ return -1;
+ }
+
+ if (migrate_type >= MAX_MIGRATE_TYPES) {
+ pr_debug("Out-of-bounds migratetype %u\n", migrate_type);
+ return -1;
+ }
+
if (use_pfn)
page = perf_sample__intval(sample, "pfn");
else
@@ -892,6 +902,11 @@ static int evsel__process_page_free_event(struct perf_sample *sample)
.order = order,
};
+ if (order >= MAX_PAGE_ORDER) {
+ pr_debug("Out-of-bounds order %u\n", order);
+ return -1;
+ }
+
if (use_pfn)
page = perf_sample__intval(sample, "pfn");
else
--
2.54.0.746.g67dd491aae-googEnsure CPU indexes parsed from sched switch and runtime events fit within the
MAX_CPUS limit to prevent out-of-bounds indexing.
Add explicit bounds checks for sample->cpu against MAX_CPUS inside
process_sched_switch_event, process_sched_runtime_event, and
timehist_sched_change_event. This prevents indexing beyond the boundaries
of the sched->curr_pid tracking array, avoiding potential memory corruption or
undefined behavior.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-sched.c | 15 +++++++++++++++
1 file changed, 15 insertions(+)
diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
index d984e58c7dbf..9d73c7043182 100644
--- a/tools/perf/builtin-sched.c
+++ b/tools/perf/builtin-sched.c
@@ -1791,6 +1791,11 @@ static int process_sched_switch_event(const struct perf_tool *tool,
u32 prev_pid = perf_sample__intval(sample, "prev_pid"),
next_pid = perf_sample__intval(sample, "next_pid");
+ if (this_cpu < 0 || this_cpu >= MAX_CPUS) {
+ pr_warning("Out-of-bound sample CPU %d. Skipping sample\n", this_cpu);
+ return 0;
+ }
+
if (sched->curr_pid[this_cpu] != (u32)-1) {
/*
* Are we trying to switch away a PID that is
@@ -1813,6 +1818,11 @@ static int process_sched_runtime_event(const struct perf_tool *tool,
{
struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
+ if (sample->cpu >= MAX_CPUS) {
+ pr_warning("Out-of-bound sample CPU %u. Skipping sample\n", sample->cpu);
+ return 0;
+ }
+
if (sched->tp_handler->runtime_event)
return sched->tp_handler->runtime_event(sched, sample, machine);
@@ -2775,6 +2785,11 @@ static int timehist_sched_change_event(const struct perf_tool *tool,
int rc = 0;
const char state = perf_sample__taskstate(sample, "prev_state");
+ if (sample->cpu >= MAX_CPUS) {
+ pr_warning("Out-of-bound sample CPU %d. Skipping sample\n", sample->cpu);
+ return 0;
+ }
+
addr_location__init(&al);
if (machine__resolve(machine, &al, sample) < 0) {
pr_err("problem processing %d event. skipping it\n",
--
2.54.0.746.g67dd491aae-googPrevent out-of-bounds writes/reads in CPU state tracking arrays by enforcing
strict MAX_CPUS bounds checks in timechart's tracepoint handlers.
Ensure that cpu_id retrieved from idle/frequency and sched tracepoints is less
than MAX_CPUS before indexing into cpus_cstate_state, cpus_cstate_start_times,
and similar tracking arrays. Also, fix an off-by-one error in the CPU iteration
loop inside end_sample_processing() by changing the loop condition from
'cpu <= tchart->numcpus' to 'cpu < tchart->numcpus'.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-timechart.c | 35 +++++++++++++++++++++++++++++++---
1 file changed, 32 insertions(+), 3 deletions(-)
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c
index d2e15d02304e..630756bebe32 100644
--- a/tools/perf/builtin-timechart.c
+++ b/tools/perf/builtin-timechart.c
@@ -605,6 +605,10 @@ process_sample_cpu_idle(struct timechart *tchart __maybe_unused,
u32 state = perf_sample__intval(sample, "state");
u32 cpu_id = perf_sample__intval(sample, "cpu_id");
+ if (cpu_id >= MAX_CPUS) {
+ pr_debug("Out-of-bounds cpu_id %u\n", cpu_id);
+ return -1;
+ }
if (state == (u32)PWR_EVENT_EXIT)
c_state_end(tchart, cpu_id, sample->time);
else
@@ -620,6 +624,10 @@ process_sample_cpu_frequency(struct timechart *tchart,
u32 state = perf_sample__intval(sample, "state");
u32 cpu_id = perf_sample__intval(sample, "cpu_id");
+ if (cpu_id >= MAX_CPUS) {
+ pr_debug("Out-of-bounds cpu_id %u\n", cpu_id);
+ return -1;
+ }
p_state_change(tchart, cpu_id, sample->time, state);
return 0;
}
@@ -633,6 +641,10 @@ process_sample_sched_wakeup(struct timechart *tchart,
int waker = perf_sample__intval(sample, "common_pid");
int wakee = perf_sample__intval(sample, "pid");
+ if (sample->cpu >= MAX_CPUS) {
+ pr_debug("Out-of-bounds cpu %u\n", sample->cpu);
+ return -1;
+ }
sched_wakeup(tchart, sample->cpu, sample->time, waker, wakee, flags, backtrace);
return 0;
}
@@ -646,6 +658,10 @@ process_sample_sched_switch(struct timechart *tchart,
int next_pid = perf_sample__intval(sample, "next_pid");
u64 prev_state = perf_sample__intval(sample, "prev_state");
+ if (sample->cpu >= MAX_CPUS) {
+ pr_debug("Out-of-bounds cpu %u\n", sample->cpu);
+ return -1;
+ }
sched_switch(tchart, sample->cpu, sample->time, prev_pid, next_pid,
prev_state, backtrace);
return 0;
@@ -660,6 +676,10 @@ process_sample_power_start(struct timechart *tchart __maybe_unused,
u64 cpu_id = perf_sample__intval(sample, "cpu_id");
u64 value = perf_sample__intval(sample, "value");
+ if (cpu_id >= MAX_CPUS) {
+ pr_debug("Out-of-bounds cpu_id %llu\n", (unsigned long long)cpu_id);
+ return -1;
+ }
c_state_start(cpu_id, sample->time, value);
return 0;
}
@@ -669,6 +689,10 @@ process_sample_power_end(struct timechart *tchart,
struct perf_sample *sample,
const char *backtrace __maybe_unused)
{
+ if (sample->cpu >= MAX_CPUS) {
+ pr_debug("Out-of-bounds cpu %u\n", sample->cpu);
+ return -1;
+ }
c_state_end(tchart, sample->cpu, sample->time);
return 0;
}
@@ -681,6 +705,10 @@ process_sample_power_frequency(struct timechart *tchart,
u64 cpu_id = perf_sample__intval(sample, "cpu_id");
u64 value = perf_sample__intval(sample, "value");
+ if (cpu_id >= MAX_CPUS) {
+ pr_debug("Out-of-bounds cpu_id %llu\n", (unsigned long long)cpu_id);
+ return -1;
+ }
p_state_change(tchart, cpu_id, sample->time, value);
return 0;
}
@@ -692,10 +720,9 @@ process_sample_power_frequency(struct timechart *tchart,
*/
static void end_sample_processing(struct timechart *tchart)
{
- u64 cpu;
- struct power_event *pwr;
+ for (u64 cpu = 0; cpu < tchart->numcpus; cpu++) {
+ struct power_event *pwr;
- for (cpu = 0; cpu <= tchart->numcpus; cpu++) {
/* C state */
#if 0
pwr = zalloc(sizeof(*pwr));
@@ -1515,6 +1542,8 @@ static int process_header(struct perf_file_section *section __maybe_unused,
switch (feat) {
case HEADER_NRCPUS:
tchart->numcpus = ph->env.nr_cpus_avail;
+ if (tchart->numcpus > MAX_CPUS)
+ tchart->numcpus = MAX_CPUS;
break;
case HEADER_CPU_TOPOLOGY:
--
2.54.0.746.g67dd491aae-googPrevent out-of-bounds memory reads when parsing corrupted or maliciously crafted
perf.data files by introducing robust bounds validation to raw data accessors.
- Add a helper out_of_bounds() to check if field offsets and sizes exceed the
sample's raw_size boundary, preventing heap read overflows.
- In perf_sample__rawptr(), properly resolve newer relative dynamic tracepoint
fields (__rel_loc) by checking the boundaries before and after reading the
dynamic field descriptor.
- Byte-swap dynamic field offsets and sizes dynamically when endianness varies,
ensuring cross-endian parsing is robust.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/util/evsel.c | 54 +++++++++++++++++++++++++++++++++++++----
1 file changed, 49 insertions(+), 5 deletions(-)
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index bb48568b8101..713a250c7374 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -3700,22 +3700,63 @@ struct tep_format_field *evsel__common_field(struct evsel *evsel, const char *na
return tp_format ? tep_find_common_field(tp_format, name) : NULL;
}
+static bool out_of_bounds(const struct tep_format_field *field, int offset, int size, u32 raw_size)
+{
+ if (offset < 0) {
+ pr_warning("Negative trace point field offset %d in %s\n",
+ offset, field->name);
+ return true;
+ }
+ if (size < 0) {
+ pr_warning("Negative trace point field size %d in %s\n",
+ size, field->name);
+ return true;
+ }
+ if ((u32)offset + (u32)size > raw_size) {
+ pr_warning("Out of bound tracepoint field (%s) offset %d size %d in %u\n",
+ field->name, offset, size, raw_size);
+ return true;
+ }
+ return false;
+}
+
void *perf_sample__rawptr(struct perf_sample *sample, const char *name)
{
struct tep_format_field *field = evsel__field(sample->evsel, name);
- int offset;
+ int offset, size;
if (!field)
return NULL;
offset = field->offset;
-
+ size = field->size;
if (field->flags & TEP_FIELD_IS_DYNAMIC) {
- offset = *(int *)(sample->raw_data + field->offset);
- offset &= 0xffff;
- if (tep_field_is_relative(field->flags))
+ int dynamic_data;
+
+ if (out_of_bounds(field, offset, 4, sample->raw_size))
+ return NULL;
+
+ dynamic_data = *(int *)(sample->raw_data + field->offset);
+
+ if (sample->evsel->needs_swap)
+ dynamic_data = bswap_32(dynamic_data);
+
+ offset = dynamic_data & 0xffff;
+ size = (dynamic_data >> 16) & 0xffff;
+
+ if (tep_field_is_relative(field->flags)) {
+ /*
+ * Newer kernel feature: Relative offsets (__rel_loc).
+ * If the relative flag is set, the parsed offset is not
+ * absolute from the start of the record. Instead, it is
+ * relative to the *end* of the dynamic field descriptor
+ * itself.
+ */
offset += field->offset + field->size;
+ }
}
+ if (out_of_bounds(field, offset, size, sample->raw_size))
+ return NULL;
return sample->raw_data + offset;
}
@@ -3726,6 +3767,9 @@ u64 format_field__intval(struct tep_format_field *field, struct perf_sample *sam
u64 value;
void *ptr = sample->raw_data + field->offset;
+ if (out_of_bounds(field, field->offset, field->size, sample->raw_size))
+ return 0;
+
switch (field->size) {
case 1:
return *(u8 *)ptr;
--
2.54.0.746.g67dd491aae-googThere is a double free in the record array due to how parse_options
will mutate the array. Fix by keeping an array that isn't
mutated. Ensure kwork_usage is freed on all paths.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-kwork.c | 30 +++++++++++++++++++++---------
1 file changed, 21 insertions(+), 9 deletions(-)
diff --git a/tools/perf/builtin-kwork.c b/tools/perf/builtin-kwork.c
index 59a54d11f7fa..a4604e152002 100644
--- a/tools/perf/builtin-kwork.c
+++ b/tools/perf/builtin-kwork.c
@@ -2258,7 +2258,7 @@ static void setup_event_list(struct perf_kwork *kwork,
static int perf_kwork__record(struct perf_kwork *kwork,
int argc, const char **argv)
{
- const char **rec_argv;
+ const char **rec_argv, **to_free = NULL;
unsigned int rec_argc, i, j;
struct kwork_class *class;
int ret;
@@ -2295,16 +2295,27 @@ static int perf_kwork__record(struct perf_kwork *kwork,
BUG_ON(i != rec_argc);
+ /* Save the pointers as the array will be mutated by cmd_record. */
+ to_free = calloc(rec_argc + 1, sizeof(char *));
+ if (to_free == NULL) {
+ ret = -ENOMEM;
+ goto EXIT;
+ }
+
pr_debug("record comm: ");
- for (j = 0; j < rec_argc; j++)
+ for (j = 0; j < rec_argc; j++) {
pr_debug("%s ", rec_argv[j]);
+ to_free[j] = rec_argv[j];
+ }
pr_debug("\n");
ret = cmd_record(i, rec_argv);
EXIT:
for (i = 0; i < rec_argc; i++)
- free((void *)rec_argv[i]);
+ free((void *)(to_free ? to_free[i] : rec_argv[i]));
+
+ free(to_free);
free(rec_argv);
return ret;
}
@@ -2447,6 +2458,7 @@ int cmd_kwork(int argc, const char **argv)
const char *const kwork_subcommands[] = {
"record", "report", "latency", "timehist", "top", NULL
};
+ int ret = 0;
perf_tool__init(&kwork.tool, /*ordered_events=*/true);
kwork.tool.mmap = perf_event__process_mmap;
@@ -2463,7 +2475,7 @@ int cmd_kwork(int argc, const char **argv)
if (strlen(argv[0]) > 2 && strstarts("record", argv[0])) {
setup_event_list(&kwork, kwork_options, kwork_usage);
- return perf_kwork__record(&kwork, argc, argv);
+ ret = perf_kwork__record(&kwork, argc, argv);
} else if (strlen(argv[0]) > 2 && strstarts("report", argv[0])) {
kwork.sort_order = default_report_sort_order;
if (argc > 1) {
@@ -2474,7 +2486,7 @@ int cmd_kwork(int argc, const char **argv)
kwork.report = KWORK_REPORT_RUNTIME;
setup_sorting(&kwork, report_options, report_usage);
setup_event_list(&kwork, kwork_options, kwork_usage);
- return perf_kwork__report(&kwork);
+ ret = perf_kwork__report(&kwork);
} else if (strlen(argv[0]) > 2 && strstarts("latency", argv[0])) {
kwork.sort_order = default_latency_sort_order;
if (argc > 1) {
@@ -2485,7 +2497,7 @@ int cmd_kwork(int argc, const char **argv)
kwork.report = KWORK_REPORT_LATENCY;
setup_sorting(&kwork, latency_options, latency_usage);
setup_event_list(&kwork, kwork_options, kwork_usage);
- return perf_kwork__report(&kwork);
+ ret = perf_kwork__report(&kwork);
} else if (strlen(argv[0]) > 2 && strstarts("timehist", argv[0])) {
if (argc > 1) {
argc = parse_options(argc, argv, timehist_options, timehist_usage, 0);
@@ -2494,7 +2506,7 @@ int cmd_kwork(int argc, const char **argv)
}
kwork.report = KWORK_REPORT_TIMEHIST;
setup_event_list(&kwork, kwork_options, kwork_usage);
- return perf_kwork__timehist(&kwork);
+ ret = perf_kwork__timehist(&kwork);
} else if (strlen(argv[0]) > 2 && strstarts("top", argv[0])) {
kwork.sort_order = default_top_sort_order;
if (argc > 1) {
@@ -2507,12 +2519,12 @@ int cmd_kwork(int argc, const char **argv)
kwork.event_list_str = "sched, irq, softirq";
setup_event_list(&kwork, kwork_options, kwork_usage);
setup_sorting(&kwork, top_options, top_usage);
- return perf_kwork__top(&kwork);
+ ret = perf_kwork__top(&kwork);
} else
usage_with_options(kwork_usage, kwork_options);
/* free usage string allocated by parse_options_subcommand */
free((void *)kwork_usage[0]);
- return 0;
+ return ret;
}
--
2.54.0.746.g67dd491aae-googThis commit addresses several memory management issues in builtin-kwork.c:
1. Implements a global cleanup function perf_kwork__exit to free all
kwork_work and kwork_atom_page objects at the end of the command.
2. Ensures all 'name' fields in struct kwork_work are malloc-ed (or NULL)
and properly freed by using strdup and zfree.
3. Fixes memory leaks in top_merge_tasks where kwork_work objects were
dropped without being freed.
4. Adds robustness with NULL checks for name fields.
5. Fixes workqueue_work_init to correctly resolve and strdup kernel
function names, preventing bad-free errors.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/builtin-kwork.c | 121 ++++++++++++++++++++++++++++--------
tools/perf/util/bpf_kwork.c | 22 +++++--
tools/perf/util/kwork.h | 2 +
3 files changed, 116 insertions(+), 29 deletions(-)
diff --git a/tools/perf/builtin-kwork.c b/tools/perf/builtin-kwork.c
index a4604e152002..f793ea578515 100644
--- a/tools/perf/builtin-kwork.c
+++ b/tools/perf/builtin-kwork.c
@@ -323,8 +323,8 @@ static struct kwork_work *work_search(struct rb_root_cached *root,
else if (cmp < 0)
node = node->rb_right;
else {
- if (work->name == NULL)
- work->name = key->name;
+ if (work->name == NULL && key->name != NULL)
+ work->name = strdup(key->name);
return work;
}
}
@@ -371,11 +371,54 @@ static struct kwork_work *work_new(struct kwork_work *key)
work->id = key->id;
work->cpu = key->cpu;
- work->name = key->name;
+ work->name = key->name ? strdup(key->name) : NULL;
work->class = key->class;
return work;
}
+
+static void work_delete(struct kwork_work *work)
+{
+ if (work) {
+ work_exit(work);
+ free(work);
+ }
+}
+
+static void kwork_work__free_root(struct rb_root_cached *root)
+{
+ struct rb_node *next;
+ struct kwork_work *work;
+
+ while ((next = rb_first_cached(root))) {
+ work = rb_entry(next, struct kwork_work, node);
+ rb_erase_cached(next, root);
+ work_delete(work);
+ }
+}
+
+static void perf_kwork__exit(struct perf_kwork *kwork)
+{
+ struct kwork_class *class;
+ struct kwork_atom_page *page, *tmp_page;
+
+ list_for_each_entry(class, &kwork->class_list, list) {
+ kwork_work__free_root(&class->work_root);
+ }
+
+ kwork_work__free_root(&kwork->sorted_work_root);
+
+ list_for_each_entry_safe(page, tmp_page, &kwork->atom_page_list, list) {
+ list_del_init(&page->list);
+ free(page);
+ }
+
+ INIT_LIST_HEAD(&kwork->class_list);
+ INIT_LIST_HEAD(&kwork->atom_page_list);
+ INIT_LIST_HEAD(&kwork->sort_list);
+ INIT_LIST_HEAD(&kwork->cmp_id);
+}
+
static struct kwork_work *work_findnew(struct rb_root_cached *root,
struct kwork_work *key,
struct list_head *sort_list)
@@ -453,25 +496,29 @@ static int work_push_atom(struct perf_kwork *kwork,
struct kwork_work **ret_work,
bool overwrite)
{
- struct kwork_atom *atom, *dst_atom, *last_atom;
+ struct kwork_atom *atom = NULL, *dst_atom, *last_atom;
struct kwork_work *work, key;
+ int ret = 0;
BUG_ON(class->work_init == NULL);
class->work_init(kwork, class, &key, src_type, sample, machine);
atom = atom_new(kwork, sample);
- if (atom == NULL)
- return -1;
+ if (atom == NULL) {
+ ret = -1;
+ goto out;
+ }
work = work_findnew(&class->work_root, &key, &kwork->cmp_id);
if (work == NULL) {
atom_free(atom);
- return -1;
+ ret = -1;
+ goto out;
}
if (!profile_event_match(kwork, work, sample)) {
atom_free(atom);
- return 0;
+ goto out;
}
if (dst_type < KWORK_TRACE_MAX) {
@@ -498,8 +545,9 @@ static int work_push_atom(struct perf_kwork *kwork,
}
list_add_tail(&atom->list, &work->atom_list[src_type]);
-
- return 0;
+out:
+ work_exit(&key);
+ return ret;
}
static struct kwork_atom *work_pop_atom(struct perf_kwork *kwork,
@@ -510,7 +558,7 @@ static struct kwork_atom *work_pop_atom(struct perf_kwork *kwork,
struct machine *machine,
struct kwork_work **ret_work)
{
- struct kwork_atom *atom, *src_atom;
+ struct kwork_atom *atom = NULL, *src_atom;
struct kwork_work *work, key;
BUG_ON(class->work_init == NULL);
@@ -521,15 +569,15 @@ static struct kwork_atom *work_pop_atom(struct perf_kwork *kwork,
*ret_work = work;
if (work == NULL)
- return NULL;
+ goto out;
if (!profile_event_match(kwork, work, sample))
- return NULL;
+ goto out;
atom = list_last_entry_or_null(&work->atom_list[dst_type],
struct kwork_atom, list);
if (atom != NULL)
- return atom;
+ goto out;
src_atom = atom_new(kwork, sample);
if (src_atom != NULL)
@@ -538,8 +586,9 @@ static struct kwork_atom *work_pop_atom(struct perf_kwork *kwork,
if (ret_work != NULL)
*ret_work = NULL;
}
-
- return NULL;
+out:
+ work_exit(&key);
+ return atom;
}
static struct kwork_work *find_work_by_id(struct rb_root_cached *root,
@@ -1002,13 +1051,16 @@ static void irq_work_init(struct perf_kwork *kwork,
work->name = NULL;
} else {
work->id = perf_sample__intval(sample, "irq");
- work->name = perf_sample__strval(sample, "name");
+ work->name = strdup(perf_sample__strval(sample, "name") ?: "<unknown>");
}
}
static void irq_work_name(struct kwork_work *work, char *buf, int len)
{
- snprintf(buf, len, "%s:%" PRIu64 "", work->name, work->id);
+ if (work->name != NULL)
+ snprintf(buf, len, "%s:%" PRIu64 "", work->name, work->id);
+ else
+ snprintf(buf, len, "%" PRIu64 "", work->id);
}
static struct kwork_class kwork_irq = {
@@ -1135,7 +1187,10 @@ static void softirq_work_init(struct perf_kwork *kwork,
static void softirq_work_name(struct kwork_work *work, char *buf, int len)
{
- snprintf(buf, len, "(s)%s:%" PRIu64 "", work->name, work->id);
+ if (work->name != NULL)
+ snprintf(buf, len, "(s)%s:%" PRIu64 "", work->name, work->id);
+ else
+ snprintf(buf, len, "(s)%" PRIu64 "", work->id);
}
static struct kwork_class kwork_softirq = {
@@ -1220,8 +1275,14 @@ static void workqueue_work_init(struct perf_kwork *kwork __maybe_unused,
work->class = class;
work->cpu = sample->cpu;
work->id = perf_sample__intval(sample, "work");
- work->name = function_addr == 0 ? NULL :
- machine__resolve_kernel_addr(machine, &function_addr, &modp);
+ work->name = NULL;
+
+ if (function_addr != 0) {
+ const char *name = machine__resolve_kernel_addr(machine, &function_addr, &modp);
+
+ if (name)
+ work->name = strdup(name);
+ }
}
static void workqueue_work_name(struct kwork_work *work, char *buf, int len)
@@ -1284,16 +1345,16 @@ static void sched_work_init(struct perf_kwork *kwork __maybe_unused,
if (src_type == KWORK_TRACE_EXIT) {
work->id = perf_sample__intval(sample, "prev_pid");
- work->name = strdup(perf_sample__strval(sample, "prev_comm"));
+ work->name = strdup(perf_sample__strval(sample, "prev_comm") ?: "<unknown>");
} else if (src_type == KWORK_TRACE_ENTRY) {
work->id = perf_sample__intval(sample, "next_pid");
- work->name = strdup(perf_sample__strval(sample, "next_comm"));
+ work->name = strdup(perf_sample__strval(sample, "next_comm") ?: "<unknown>");
}
}
static void sched_work_name(struct kwork_work *work, char *buf, int len)
{
- snprintf(buf, len, "%s", work->name);
+ snprintf(buf, len, "%s", work->name ?: "");
}
static struct kwork_class kwork_sched = {
@@ -2100,8 +2161,10 @@ static void top_merge_tasks(struct perf_kwork *kwork)
rb_erase_cached(node, &class->work_root);
data = rb_entry(node, struct kwork_work, node);
- if (!profile_name_match(kwork, data))
+ if (!profile_name_match(kwork, data)) {
+ work_delete(data);
continue;
+ }
cpu = data->cpu;
merged_work = find_work_by_id(&merged_root, data->id,
@@ -2109,11 +2172,17 @@ static void top_merge_tasks(struct perf_kwork *kwork)
if (!merged_work) {
work_insert(&merged_root, data, &kwork->cmp_id);
} else {
+ if (merged_work->name == NULL && data->name != NULL)
+ merged_work->name = strdup(data->name);
+
merged_work->total_runtime += data->total_runtime;
merged_work->cpu_usage += data->cpu_usage;
}
top_calc_load_runtime(kwork, data);
+
+ if (merged_work)
+ work_delete(data);
}
work_sort(kwork, class, &merged_root);
@@ -2523,6 +2592,8 @@ int cmd_kwork(int argc, const char **argv)
} else
usage_with_options(kwork_usage, kwork_options);
+ perf_kwork__exit(&kwork);
+
/* free usage string allocated by parse_options_subcommand */
free((void *)kwork_usage[0]);
diff --git a/tools/perf/util/bpf_kwork.c b/tools/perf/util/bpf_kwork.c
index d3a2e548f2b6..70b711e1a23f 100644
--- a/tools/perf/util/bpf_kwork.c
+++ b/tools/perf/util/bpf_kwork.c
@@ -53,6 +53,14 @@ static struct kwork_trace_bpf *skel;
static struct timespec ts_start;
static struct timespec ts_end;
+void work_exit(struct kwork_work *work)
+{
+ if (work) {
+ free(work->name);
+ work->name = NULL;
+ }
+}
+
void perf_kwork__trace_start(void)
{
clock_gettime(CLOCK_MONOTONIC, &ts_start);
@@ -273,6 +281,7 @@ static int add_work(struct perf_kwork *kwork,
.cpu = key->cpu,
};
enum kwork_class_type type = key->type;
+ int ret = 0;
if (!valid_kwork_class_type(type)) {
pr_debug("Invalid class type %d to add work\n", type);
@@ -287,8 +296,10 @@ static int add_work(struct perf_kwork *kwork,
return -1;
work = kwork->add_work(kwork, tmp.class, &tmp);
- if (work == NULL)
- return -1;
+ if (work == NULL) {
+ ret = -1;
+ goto out;
+ }
if (kwork->report == KWORK_REPORT_RUNTIME) {
work->nr_atoms = data->nr;
@@ -304,13 +315,16 @@ static int add_work(struct perf_kwork *kwork,
work->max_latency_end = data->max_time_end;
} else {
pr_debug("Invalid bpf report type %d\n", kwork->report);
- return -1;
+ ret = -1;
+ goto out;
}
kwork->timestart = (u64)ts_start.tv_sec * NSEC_PER_SEC + ts_start.tv_nsec;
kwork->timeend = (u64)ts_end.tv_sec * NSEC_PER_SEC + ts_end.tv_nsec;
- return 0;
+out:
+ work_exit(&tmp);
+ return ret;
}
int perf_kwork__report_read_bpf(struct perf_kwork *kwork)
diff --git a/tools/perf/util/kwork.h b/tools/perf/util/kwork.h
index abf637d44794..c96f388b3159 100644
--- a/tools/perf/util/kwork.h
+++ b/tools/perf/util/kwork.h
@@ -164,6 +164,8 @@ struct kwork_class {
char *buf, int len);
};
+void work_exit(struct kwork_work *work);
+
struct trace_kwork_handler {
int (*raise_event)(struct perf_kwork *kwork,
struct kwork_class *class,
--
2.54.0.746.g67dd491aae-goog
Hello,
kernel test robot noticed "perf-stat-tests.perf.make.fail" on:
commit: 7f9aae3554b369864c01505b2ec02ddf6f6e12f3 ("[PATCH v14 32/32] perf kwork: Fix memory management of kwork_work")
url: https://github.com/intel-lab-lkp/linux/commits/Ian-Rogers/perf-tool-Remove-evsel-from-tool-APIs-that-pass-the-sample/20260521-061855
patch link: https://lore.kernel.org/all/20260520190538.142018-33-irogers@google.com/
patch subject: [PATCH v14 32/32] perf kwork: Fix memory management of kwork_work
in testcase: perf-stat-tests
version:
with following parameters:
config: x86_64-rhel-9.4-bpf
compiler: gcc-14
test machine: 16 threads Intel(R) Core(TM) i7-13620H (Raptor Lake) with 32G memory
(please refer to attached dmesg/kmsg for entire log/backtrace)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <oliver.sang@intel.com>
| Closes: https://lore.kernel.org/oe-lkp/202605251514.4b99efea-lkp@intel.com
The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20260525/202605251514.4b99efea-lkp@intel.com
==> /tmp/stderr <==
Makefile.config:661: Warning: Disabled BPF skeletons as clang (clang) is missing
Makefile.config:1134: Rust is not found. Test workloads with rust are disabled.
PERF_VERSION = 7.1.0-rc2
/usr/bin/ld: perf-in.o: in function `work_delete':
/usr/src/perf_selftests-x86_64-rhel-9.4-bpf-7f9aae3554b369864c01505b2ec02ddf6f6e12f3/tools/perf/builtin-kwork.c:383:(.text+0x1b569b): undefined reference to `work_exit'
/usr/bin/ld: perf-in.o: in function `work_push_atom':
/usr/src/perf_selftests-x86_64-rhel-9.4-bpf-7f9aae3554b369864c01505b2ec02ddf6f6e12f3/tools/perf/builtin-kwork.c:549:(.text+0x1b77d3): undefined reference to `work_exit'
/usr/bin/ld: perf-in.o: in function `work_pop_atom':
/usr/src/perf_selftests-x86_64-rhel-9.4-bpf-7f9aae3554b369864c01505b2ec02ddf6f6e12f3/tools/perf/builtin-kwork.c:590:(.text+0x1b7fc6): undefined reference to `work_exit'
collect2: error: ld returned 1 exit status
make[2]: *** [Makefile.perf:803: perf] Error 1
make[2]: *** Waiting for unfinished jobs....
make[1]: *** [Makefile.perf:289: sub-make] Error 2
make: *** [Makefile:76: all] Error 2
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
© 2016 - 2026 Red Hat, Inc.