Re: [PATCH v1 1/6] perf parse-events: Fixes relating to no_value terms

[Arnaldo Carvalho de Melo <acme@kernel.org>]

Em Tue, Sep 05, 2023 at 10:39:38AM +0100, James Clark escreveu:
> 
> 
> On 02/09/2023 00:39, Ian Rogers wrote:
> > A term may have no value in which case it is assumed to have a value
> > of 1. It doesn't just apply to alias/event terms so change the
> > parse_events_term__to_strbuf assert.
> > 
> > Commit 99e7138eb789 ("perf tools: Fail on using multiple bits long
> > terms without value") made it so that no_value terms could only be for
> > a single bit. Prior to commit 64199ae4b8a3 ("perf parse-events: Fix
> > propagation of term's no_value when cloning") this missed a test case
> > where config1 had no_value.
> > 
> > Fixes: 64199ae4b8a3 ("perf parse-events: Fix propagation of term's no_value when cloning")
> > Signed-off-by: Ian Rogers <irogers@google.com>
> 
> For the whole set:
> 
> Reviewed-by: James Clark <james.clark@arm.com>

Thanks for providing it, in the future please consider replying to the
cover letter (0/6) to do that, so that b4 can collect it, doing it for
the 1st patch it will collect just for it, not for the whole set as you
intended.

- Arnaldo
 
> > ---
> >  tools/perf/tests/parse-events.c | 2 +-
> >  tools/perf/util/parse-events.c  | 2 +-
> >  tools/perf/util/parse-events.h  | 4 ++--
> >  3 files changed, 4 insertions(+), 4 deletions(-)
> > 
> > diff --git a/tools/perf/tests/parse-events.c b/tools/perf/tests/parse-events.c
> > index d86076d575ed..d47f1f871164 100644
> > --- a/tools/perf/tests/parse-events.c
> > +++ b/tools/perf/tests/parse-events.c
> > @@ -2170,7 +2170,7 @@ static const struct evlist_test test__events[] = {
> >  
> >  static const struct evlist_test test__events_pmu[] = {
> >  	{
> > -		.name  = "cpu/config=10,config1,config2=3,period=1000/u",
> > +		.name  = "cpu/config=10,config1=1,config2=3,period=1000/u",
> >  		.valid = test__pmu_cpu_valid,
> >  		.check = test__checkevent_pmu,
> >  		/* 0 */
> > diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
> > index 68fe2c4ff49f..65608a3cba81 100644
> > --- a/tools/perf/util/parse-events.c
> > +++ b/tools/perf/util/parse-events.c
> > @@ -2607,7 +2607,7 @@ int parse_events_term__to_strbuf(struct list_head *term_list, struct strbuf *sb)
> >  
> >  		if (term->type_val == PARSE_EVENTS__TERM_TYPE_NUM)
> >  			if (term->no_value) {
> > -				assert(term->type_term == PARSE_EVENTS__TERM_TYPE_USER);
> > +				assert(term->val.num == 1);
> >  				ret = strbuf_addf(sb, "%s", term->config);
> >  			} else
> >  				ret = strbuf_addf(sb, "%s=%#"PRIx64, term->config, term->val.num);
> > diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h
> > index 855b0725c5d4..594e5d2dc67f 100644
> > --- a/tools/perf/util/parse-events.h
> > +++ b/tools/perf/util/parse-events.h
> > @@ -124,8 +124,8 @@ struct parse_events_term {
> >  	 */
> >  	bool weak;
> >  	/**
> > -	 * @no_value: Is there no value. TODO: this should really be part of
> > -	 * type_val.
> > +	 * @no_value: Is there no value. If a numeric term has no value then the
> > +	 * value is assumed to be 1. An event name also has no value.
> >  	 */
> >  	bool no_value;
> >  };

-- 

- Arnaldo