Re: [PATCH 2/3] perf build: Use libtraceevent from the system

From: Ian Rogers
Date: Thu Dec 08 2022 - 18:05:50 EST

Next message: Eric Whitney: "Re: [PATCH v4 3/3] ext4: add check pending tree when evict inode"
Previous message: Jiri Olsa: "Re: BUG: unable to handle kernel paging request in bpf_dispatcher_xdp"
In reply to: Ian Rogers: "Re: [PATCH 2/3] perf build: Use libtraceevent from the system"
Next in thread: Arnaldo Carvalho de Melo: "Re: [PATCH 2/3] perf build: Use libtraceevent from the system"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Thu, Dec 8, 2022 at 3:00 PM Ian Rogers <irogers@xxxxxxxxxx> wrote:
>
> On Thu, Dec 8, 2022 at 2:32 PM Arnaldo Carvalho de Melo <acme@xxxxxxxxxx> wrote:
> >
> > Em Thu, Dec 08, 2022 at 07:04:52PM -0300, Arnaldo Carvalho de Melo escreveu:
> > > Em Thu, Dec 08, 2022 at 12:21:20PM +0530, Athira Rajeev escreveu:
> > > > > On 07-Dec-2022, at 10:57 PM, Arnaldo Carvalho de Melo <acme@xxxxxxxxxx> wrote:
> > > > > Can you try again? tmp.perf/core? That "tmp." part means its a force
> > > > > pushed branch, so I just force pushed with some arch specific fixes, now
> > > > > I'm down to (removing the successful builds and unrelated failures, now
> > > > > related to libbpf's F_DUPFD_CLOEXEC kaboom):
> > >
> > > > Ok Arnaldo, Sure, I will check with updated branch
> > >
> > > > > 5 7.38 fedora:34 : FAIL gcc version 11.3.1 20220421 (Red Hat 11.3.1-2) (GCC)
> > > > > /git/perf-6.1.0-rc6/tools/perf/util/evsel.c: In function ‘evsel__rawptr’:
> > > > > /git/perf-6.1.0-rc6/tools/perf/util/evsel.c:2787:36: error: ‘TEP_FIELD_IS_RELATIVE’ undeclared (first use in this function); did you mean ‘TEP_FIELD_IS_FLAG’?
> > > > > 2787 | if (field->flags & TEP_FIELD_IS_RELATIVE)
> > > > > | ^~~~~~~~~~~~~~~~~~~~~
> > > > > | TEP_FIELD_IS_FLAG
> > >
> > > > I observed same issue as updated here:
> > > > https://lore.kernel.org/lkml/10476A85-3F75-4C91-AB5B-E5B136F31297@xxxxxxxxxxxxxxxxxx/
> > >
> > > > Looks like TEP_FIELD_IS_RELATIVE is not defined in header file of the system installed version.
> > > > whereas it is there in header file in tools/lib/traceevent
> > >
> > > > # grep TEP_FIELD_IS_RELATIVE /usr/include/traceevent/event-parse.h
> > > > # grep TEP_FIELD_IS_RELATIVE ../lib/traceevent/event-parse.h
> > > > TEP_FIELD_IS_RELATIVE = 256,
> > >
> > > Right, I had noticed that as well, so as a prep patch I'm adding the
> > > patch below, before Ian's. Please check and provide an
> > > Acked-by/Tested-by/Reviewed-by if possible.
> >
> > I ended up with the one below, _after_ Ian's patch as I had some trouble grafting
> > it before and had already tested it this way multiple times, I'm pushing
> > this to tmp/perf.core.
> >
> > - Arnaldo
> >
> > commit a0ec3ef4590e7e82339b6f318e671a6b69fd94e8
> > Author: Arnaldo Carvalho de Melo <acme@xxxxxxxxxx>
> > Date: Thu Dec 8 19:28:58 2022 -0300
> >
> > perf tools: Add test to check if libtracevent has TEP_FIELD_IS_RELATIVE
> >
>
> Could HAVE_LIBTRACEEVENT_TEP_FIELD_IS_RELATIVE be a version rather
> than feature check, something like:
>
> #define MAKE_LIBTRACEEVENT_VERSION(maj, mid, min) (maj * 255 *255 +
> mid * 255 + min)
> #define HAVE_LIBTRACEEVENT_TEP_FIELD_IS_RELATIVE
> (LIBTRACEEVENT_VERSION >= MAKE_LIBTRACEEVENT_VERSION(1, 5, 0)
>
> Note, we have LIBTRACEEVENT_VERSION computed in Makefile.config by way
> of pkg-config.
>
> Thanks,
> Ian

LIBTRACEEVENT_VERSION had a workaround if pkg-config wasn't present to
default to 1.1.0 for its use in:
https://lore.kernel.org/linux-perf-users/20210923001024.550263-4-irogers@xxxxxxxxxx/
We should un-revert that change now the -Itools/lib issue is resolved
(1.1 traceevent headers compiled against newer system libtraceevent
library).

Thanks,
Ian

> > Some distros have older versions of libtraceevent where
> > TEP_FIELD_IS_RELATIVE and its associated semantics are not present, so
> > we need to test for its presence and cope when it isn't present when
> > building with LIBTRACEEVENT_DYNAMIC=1.
> >
> > Reported-by: Athira Jajeev <atrajeev@xxxxxxxxxxxxxxxxxx>
> > Cc: Adrian Hunter <adrian.hunter@xxxxxxxxx>
> > Cc: Alexander Shishkin <alexander.shishkin@xxxxxxxxxxxxxxx>
> > Cc: Ian Rogers <irogers@xxxxxxxxxx>
> > Cc: Ingo Molnar <mingo@xxxxxxxxxx>
> > Cc: Jiri Olsa <jolsa@xxxxxxxxxx>
> > Cc: Mark Rutland <mark.rutland@xxxxxxx>
> > Cc: Namhyung Kim <namhyung@xxxxxxxxxx>
> > Cc: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
> > Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>,
> > Cc: Stephane Eranian <eranian@xxxxxxxxxx>
> > Signed-off-by: Arnaldo Carvalho de Melo <acme@xxxxxxxxxx>
> >
> > diff --git a/tools/build/Makefile.feature b/tools/build/Makefile.feature
> > index 38f8851bd7cbdb0e..c6c9b7c7723d4290 100644
> > --- a/tools/build/Makefile.feature
> > +++ b/tools/build/Makefile.feature
> > @@ -71,7 +71,8 @@ FEATURE_TESTS_BASIC := \
> > libzstd \
> > disassembler-four-args \
> > disassembler-init-styled \
> > - file-handle
> > + file-handle \
> > + libtraceevent-tep_field_is_relative
> >
> > # FEATURE_TESTS_BASIC + FEATURE_TESTS_EXTRA is the complete list
> > # of all feature tests
> > diff --git a/tools/build/feature/Makefile b/tools/build/feature/Makefile
> > index 690fe97be1904399..24246df35ac30eeb 100644
> > --- a/tools/build/feature/Makefile
> > +++ b/tools/build/feature/Makefile
> > @@ -81,7 +81,8 @@ FILES= \
> > test-libzstd.bin \
> > test-clang-bpf-co-re.bin \
> > test-file-handle.bin \
> > - test-libpfm4.bin
> > + test-libpfm4.bin \
> > + test-libtraceevent-tep_field_is_relative
> >
> > FILES := $(addprefix $(OUTPUT),$(FILES))
> >
> > @@ -159,6 +160,9 @@ $(OUTPUT)test-dwarf_getlocations.bin:
> > $(OUTPUT)test-libelf-getphdrnum.bin:
> > $(BUILD) -lelf
> >
> > +$(OUTPUT)test-libtraceevent-tep_field_is_relative.bin:
> > + $(BUILD) -ltraceevent
> > +
> > $(OUTPUT)test-libelf-gelf_getnote.bin:
> > $(BUILD) -lelf
> >
> > diff --git a/tools/perf/Makefile.config b/tools/perf/Makefile.config
> > index 680228e19c1aabdb..e29975664f06bd61 100644
> > --- a/tools/perf/Makefile.config
> > +++ b/tools/perf/Makefile.config
> > @@ -1197,6 +1197,10 @@ ifneq ($(NO_LIBTRACEEVENT),1)
> > LIBTRACEEVENT_VERSION_CPP := $(shell expr $(LIBTRACEEVENT_VERSION_1) \* 255 \* 255 + $(LIBTRACEEVENT_VERSION_2) \* 255 + $(LIBTRACEEVENT_VERSION_3))
> > CFLAGS += -DLIBTRACEEVENT_VERSION=$(LIBTRACEEVENT_VERSION_CPP)
> > $(call detected,CONFIG_LIBTRACEEVENT)
> > + $(call feature_check,libtraceevent-tep_field_is_relative)
> > + ifeq ($(feature-libtraceevent-tep_field_is_relative), 1)
> > + CFLAGS += -DHAVE_LIBTRACEEVENT_TEP_FIELD_IS_RELATIVE
> > + endif
> > else
> > dummy := $(warning Warning: libtraceevent is missing limiting functionality, please install libtraceevent-dev/libtraceevent-devel)
> > endif
> > diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
> > index 6909cd9f48d11644..c9802ab549d7782f 100644
> > --- a/tools/perf/builtin-trace.c
> > +++ b/tools/perf/builtin-trace.c
> > @@ -2729,8 +2729,10 @@ static size_t trace__fprintf_tp_fields(struct trace *trace, struct evsel *evsel,
> > offset = format_field__intval(field, sample, evsel->needs_swap);
> > syscall_arg.len = offset >> 16;
> > offset &= 0xffff;
> > +#ifdef HAVE_LIBTRACEEVENT_TEP_FIELD_IS_RELATIVE
> > if (field->flags & TEP_FIELD_IS_RELATIVE)
> > offset += field->offset + field->size;
> > +#endif
> > }
> >
> > val = (uintptr_t)(sample->raw_data + offset);
> > diff --git a/tools/perf/util/data-convert-bt.c b/tools/perf/util/data-convert-bt.c
> > index 8031b586e81384e7..b842273458b862bd 100644
> > --- a/tools/perf/util/data-convert-bt.c
> > +++ b/tools/perf/util/data-convert-bt.c
> > @@ -322,8 +322,10 @@ static int add_tracepoint_field_value(struct ctf_writer *cw,
> > offset = tmp_val;
> > len = offset >> 16;
> > offset &= 0xffff;
> > +#ifdef HAVE_LIBTRACEEVENT_TEP_FIELD_IS_RELATIVE
> > if (flags & TEP_FIELD_IS_RELATIVE)
> > offset += fmtf->offset + fmtf->size;
> > +#endif
> > }
> >
> > if (flags & TEP_FIELD_IS_ARRAY) {
> > diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
> > index ca911856c4b1d3c8..77b2cf5a214ec1e4 100644
> > --- a/tools/perf/util/evsel.c
> > +++ b/tools/perf/util/evsel.c
> > @@ -2784,8 +2784,10 @@ void *evsel__rawptr(struct evsel *evsel, struct perf_sample *sample, const char
> > if (field->flags & TEP_FIELD_IS_DYNAMIC) {
> > offset = *(int *)(sample->raw_data + field->offset);
> > offset &= 0xffff;
> > +#ifdef HAVE_LIBTRACEEVENT_TEP_FIELD_IS_RELATIVE
> > if (field->flags & TEP_FIELD_IS_RELATIVE)
> > offset += field->offset + field->size;
> > +#endif
> > }
> >
> > return sample->raw_data + offset;
> > diff --git a/tools/perf/util/python.c b/tools/perf/util/python.c
> > index 6fb84b7455b8392d..7320f7f777fe0e56 100644
> > --- a/tools/perf/util/python.c
> > +++ b/tools/perf/util/python.c
> > @@ -442,8 +442,10 @@ tracepoint_field(struct pyrf_event *pe, struct tep_format_field *field)
> > offset = val;
> > len = offset >> 16;
> > offset &= 0xffff;
> > +#ifdef HAVE_LIBTRACEEVENT_TEP_FIELD_IS_RELATIVE
> > if (field->flags & TEP_FIELD_IS_RELATIVE)
> > offset += field->offset + field->size;
> > +#endif
> > }
> > if (field->flags & TEP_FIELD_IS_STRING &&
> > is_printable_array(data + offset, len)) {
> > diff --git a/tools/perf/util/scripting-engines/trace-event-perl.c b/tools/perf/util/scripting-engines/trace-event-perl.c
> > index 0bacb49408f84adf..c097b7934fd4ebe9 100644
> > --- a/tools/perf/util/scripting-engines/trace-event-perl.c
> > +++ b/tools/perf/util/scripting-engines/trace-event-perl.c
> > @@ -393,8 +393,10 @@ static void perl_process_tracepoint(struct perf_sample *sample,
> > if (field->flags & TEP_FIELD_IS_DYNAMIC) {
> > offset = *(int *)(data + field->offset);
> > offset &= 0xffff;
> > +#ifdef HAVE_LIBTRACEEVENT_TEP_FIELD_IS_RELATIVE
> > if (field->flags & TEP_FIELD_IS_RELATIVE)
> > offset += field->offset + field->size;
> > +#endif
> > } else
> > offset = field->offset;
> > XPUSHs(sv_2mortal(newSVpv((char *)data + offset, 0)));
> > diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c
> > index fabba21919b8fa51..e930f5f1f36d2831 100644
> > --- a/tools/perf/util/scripting-engines/trace-event-python.c
> > +++ b/tools/perf/util/scripting-engines/trace-event-python.c
> > @@ -994,8 +994,10 @@ static void python_process_tracepoint(struct perf_sample *sample,
> > offset = val;
> > len = offset >> 16;
> > offset &= 0xffff;
> > +#ifdef HAVE_LIBTRACEEVENT_TEP_FIELD_IS_RELATIVE
> > if (field->flags & TEP_FIELD_IS_RELATIVE)
> > offset += field->offset + field->size;
> > +#endif
> > }
> > if (field->flags & TEP_FIELD_IS_STRING &&
> > is_printable_array(data + offset, len)) {
> > diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c
> > index c7a97b33e13477e1..0ecc2cb137920d2d 100644
> > --- a/tools/perf/util/sort.c
> > +++ b/tools/perf/util/sort.c
> > @@ -2568,9 +2568,10 @@ static int64_t __sort__hde_cmp(struct perf_hpp_fmt *fmt,
> > tep_read_number_field(field, a->raw_data, &dyn);
> > offset = dyn & 0xffff;
> > size = (dyn >> 16) & 0xffff;
> > +#ifdef HAVE_LIBTRACEEVENT_TEP_FIELD_IS_RELATIVE
> > if (field->flags & TEP_FIELD_IS_RELATIVE)
> > offset += field->offset + field->size;
> > -
> > +#endif
> > /* record max width for output */
> > if (size > hde->dynamic_len)
> > hde->dynamic_len = size;

Next message: Eric Whitney: "Re: [PATCH v4 3/3] ext4: add check pending tree when evict inode"
Previous message: Jiri Olsa: "Re: BUG: unable to handle kernel paging request in bpf_dispatcher_xdp"
In reply to: Ian Rogers: "Re: [PATCH 2/3] perf build: Use libtraceevent from the system"
Next in thread: Arnaldo Carvalho de Melo: "Re: [PATCH 2/3] perf build: Use libtraceevent from the system"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]