From: Namhyung Kim <namhyung@kernel.org>
To: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Ian Rogers <irogers@google.com>,
Kan Liang <kan.liang@linux.intel.com>,
Jiri Olsa <jolsa@kernel.org>,
Adrian Hunter <adrian.hunter@intel.com>,
Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@kernel.org>,
LKML <linux-kernel@vger.kernel.org>,
linux-perf-users@vger.kernel.org
Subject: Re: [PATCH 1/4] perf hist: Factor out __hpp__fmt_print()
Date: Tue, 4 Jun 2024 17:42:50 -0700 [thread overview]
Message-ID: <Zl-0igwvs8EuzOMs@google.com> (raw)
In-Reply-To: <Zl8tvVPEfLM8_T_z@x1>
On Tue, Jun 04, 2024 at 12:07:41PM -0300, Arnaldo Carvalho de Melo wrote:
> On Mon, Jun 03, 2024 at 03:44:09PM -0700, Namhyung Kim wrote:
> > Split the logic to print the histogram values according to the format
> > string. This was used in 3 different places so it's better to move out
> > the logic into a function.
>
> But are all really equivalent? I had difficulty following it, perhaps it
> would be better to introduce the function and then go on making one by
> one use it instead of doing it all at once?
Sorry, I agree the code looks confusing. But they are equivalent and
just to print the value in a different way.
>
> So in the end we have:
>
> static int __hpp__fmt_print(struct perf_hpp *hpp, struct hists *hists, u64 val,
> int nr_samples, const char *fmt, int len,
> hpp_snprint_fn print_fn, enum perf_hpp_fmt_type fmtype)
> {
> if (fmtype == PERF_HPP_FMT_TYPE__PERCENT) {
> double percent = 0.0;
> u64 total = hists__total_period(hists);
>
> if (total)
> percent = 100.0 * val / total;
>
> return hpp__call_print_fn(hpp, print_fn, fmt, len, percent);
> }
> if (fmtype == PERF_HPP_FMT_TYPE__AVERAGE) {
> double avg = nr_samples ? (1.0 * val / nr_samples) : 0;
>
> return hpp__call_print_fn(hpp, print_fn, fmt, len, avg);
> }
>
> return hpp__call_print_fn(hpp, print_fn, fmt, len, val);
> }
>
> I.e. we do something with the last arg for hpp__call_print_fn() for the
> PERF_HPP_FMT_TYPE__PERCENT and PERF_HPP_FMT_TYPE__AVERAGE fmtype's, but
> we don't check for those fmtypes on each of the places that we now call
> __hpp__fmt_print(), right?
>
> That while (idx_delta--) cases is only interested in !=PERF_HPP_FMT_TYPE__RAW...
>
> Nah, after trying to describe the confusion I think I clarified it,
> probably leave it as you did :-\
:)
The 'while (idx_delta--)' part is just to fill the gap for entries
without samples. The value is always 0, so it only cares the data type
(double or u64) in the format string. RAW needs u64 while PERCENT and
AVERAGE need double.
Thanks,
Namhyung
> > +++ b/tools/perf/ui/hist.c
> > @@ -23,35 +23,42 @@
> > __ret; \
> > })
> >
> > -static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he,
> > - hpp_field_fn get_field, const char *fmt, int len,
> > - hpp_snprint_fn print_fn, enum perf_hpp_fmt_type fmtype)
> > +static int __hpp__fmt_print(struct perf_hpp *hpp, struct hists *hists, u64 val,
> > + int nr_samples, const char *fmt, int len,
> > + hpp_snprint_fn print_fn, enum perf_hpp_fmt_type fmtype)
> > {
> > - int ret;
> > - struct hists *hists = he->hists;
> > - struct evsel *evsel = hists_to_evsel(hists);
> > - char *buf = hpp->buf;
> > - size_t size = hpp->size;
> > -
> > if (fmtype == PERF_HPP_FMT_TYPE__PERCENT) {
> > double percent = 0.0;
> > u64 total = hists__total_period(hists);
> >
> > if (total)
> > - percent = 100.0 * get_field(he) / total;
> > + percent = 100.0 * val / total;
> >
> > - ret = hpp__call_print_fn(hpp, print_fn, fmt, len, percent);
> > - } else if (fmtype == PERF_HPP_FMT_TYPE__AVERAGE) {
> > - double average = 0;
> > + return hpp__call_print_fn(hpp, print_fn, fmt, len, percent);
> > + }
> >
> > - if (he->stat.nr_events)
> > - average = 1.0 * get_field(he) / he->stat.nr_events;
> > + if (fmtype == PERF_HPP_FMT_TYPE__AVERAGE) {
> > + double avg = nr_samples ? (1.0 * val / nr_samples) : 0;
> >
> > - ret = hpp__call_print_fn(hpp, print_fn, fmt, len, average);
> > - } else {
> > - ret = hpp__call_print_fn(hpp, print_fn, fmt, len, get_field(he));
> > + return hpp__call_print_fn(hpp, print_fn, fmt, len, avg);
> > }
> >
> > + return hpp__call_print_fn(hpp, print_fn, fmt, len, val);
> > +}
> > +
> > +static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he,
> > + hpp_field_fn get_field, const char *fmt, int len,
> > + hpp_snprint_fn print_fn, enum perf_hpp_fmt_type fmtype)
> > +{
> > + int ret;
> > + struct hists *hists = he->hists;
> > + struct evsel *evsel = hists_to_evsel(hists);
> > + char *buf = hpp->buf;
> > + size_t size = hpp->size;
> > +
> > + ret = __hpp__fmt_print(hpp, hists, get_field(he), he->stat.nr_events,
> > + fmt, len, print_fn, fmtype);
> > +
> > if (evsel__is_group_event(evsel)) {
> > int prev_idx, idx_delta;
> > struct hist_entry *pair;
> > @@ -72,30 +79,16 @@ static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he,
> >
> > while (idx_delta--) {
> > /*
> > - * zero-fill group members in the middle which
> > - * have no sample
> > + * zero-fill group members in the middle which have
> > + * no samples, pair->hists is not correct but it's
> > + * fine since the value is 0.
> > */
> > - if (fmtype != PERF_HPP_FMT_TYPE__RAW) {
> > - ret += hpp__call_print_fn(hpp, print_fn,
> > - fmt, len, 0.0);
> > - } else {
> > - ret += hpp__call_print_fn(hpp, print_fn,
> > - fmt, len, 0ULL);
> > - }
> > + ret += __hpp__fmt_print(hpp, pair->hists, 0, 0,
> > + fmt, len, print_fn, fmtype);
> > }
> >
> > - if (fmtype == PERF_HPP_FMT_TYPE__PERCENT) {
> > - ret += hpp__call_print_fn(hpp, print_fn, fmt, len,
> > - 100.0 * period / total);
> > - } else if (fmtype == PERF_HPP_FMT_TYPE__AVERAGE) {
> > - double avg = nr_samples ? (period / nr_samples) : 0;
> > -
> > - ret += hpp__call_print_fn(hpp, print_fn, fmt,
> > - len, avg);
> > - } else {
> > - ret += hpp__call_print_fn(hpp, print_fn, fmt,
> > - len, period);
> > - }
> > + ret += __hpp__fmt_print(hpp, pair->hists, period, nr_samples,
> > + fmt, len, print_fn, fmtype);
> >
> > prev_idx = evsel__group_idx(evsel);
> > }
> > @@ -104,15 +97,11 @@ static int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he,
> >
> > while (idx_delta--) {
> > /*
> > - * zero-fill group members at last which have no sample
> > + * zero-fill group members at last which have no sample.
> > + * the hists is not correct but it's fine like above.
> > */
> > - if (fmtype != PERF_HPP_FMT_TYPE__RAW) {
> > - ret += hpp__call_print_fn(hpp, print_fn,
> > - fmt, len, 0.0);
> > - } else {
> > - ret += hpp__call_print_fn(hpp, print_fn,
> > - fmt, len, 0ULL);
> > - }
> > + ret += __hpp__fmt_print(hpp, evsel__hists(evsel), 0, 0,
> > + fmt, len, print_fn, fmtype);
> > }
> > }
> >
> > --
> > 2.45.1.288.g0e0cd299f1-goog
next prev parent reply other threads:[~2024-06-05 0:42 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-06-03 22:44 [PATCH 0/4] perf report: Omit dummy events in the output (v2) Namhyung Kim
2024-06-03 22:44 ` [PATCH 1/4] perf hist: Factor out __hpp__fmt_print() Namhyung Kim
2024-06-04 15:07 ` Arnaldo Carvalho de Melo
2024-06-05 0:42 ` Namhyung Kim [this message]
2024-06-03 22:44 ` [PATCH 2/4] perf hist: Simplify __hpp_fmt() using hpp_fmt_data Namhyung Kim
2024-06-04 15:18 ` Arnaldo Carvalho de Melo
2024-06-05 0:47 ` Namhyung Kim
2024-06-03 22:44 ` [PATCH 3/4] perf hist: Add symbol_conf.skip_empty Namhyung Kim
2024-06-03 22:44 ` [PATCH 4/4] perf hist: Honor symbol_conf.skip_empty Namhyung Kim
2024-06-04 15:24 ` [PATCH 0/4] perf report: Omit dummy events in the output (v2) Arnaldo Carvalho de Melo
-- strict thread matches above, loose matches on Subject: below --
2024-06-07 20:29 [PATCH 0/4] perf report: Omit dummy events in the output (v3) Namhyung Kim
2024-06-07 20:29 ` [PATCH 1/4] perf hist: Factor out __hpp__fmt_print() Namhyung Kim
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=Zl-0igwvs8EuzOMs@google.com \
--to=namhyung@kernel.org \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=irogers@google.com \
--cc=jolsa@kernel.org \
--cc=kan.liang@linux.intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=peterz@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).