Re: [PATCH v1 2/3] perf parse-events: Make common term list to strbuf helper

From: Ian Rogers
Date: Wed Aug 30 2023 - 15:04:09 EST

Next message: Hans Verkuil: "Re: [PATCH v5 02/10] media: videobuf2: Access vb2_queue bufs array through helper functions"
Previous message: Marco Elver: "Re: [PATCH 11/15] stackdepot: use read/write lock"
In reply to: Ian Rogers: "Re: [PATCH v1 2/3] perf parse-events: Make common term list to strbuf helper"
Next in thread: Liang, Kan: "Re: [PATCH v1 2/3] perf parse-events: Make common term list to strbuf helper"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Wed, Aug 30, 2023 at 7:40 AM Ian Rogers <irogers@xxxxxxxxxx> wrote:
>
> On Wed, Aug 30, 2023 at 7:15 AM Liang, Kan <kan.liang@xxxxxxxxxxxxxxx> wrote:
> >
> >
> >
> > On 2023-08-30 3:07 a.m., Ian Rogers wrote:
> > > A term list is turned into a string for debug output and for the str
> > > value in the alias. Add a helper to do this based on existing code,
> > > but then fix for situations like events being identified. Use strbuf
> > > to manage the dynamic memory allocation and remove the 256 byte
> > > limit. Use in various places the string of the term list is required.
> > >
> > > Before:
> > > ```
> > > $ sudo perf stat -vv -e inst_retired.any true
> > > Using CPUID GenuineIntel-6-8D-1
> > > intel_pt default config: tsc,mtc,mtc_period=3,psb_period=3,pt,branch
> > > Attempting to add event pmu 'cpu' with 'inst_retired.any,' that may result in non-fatal errors
> > > After aliases, add event pmu 'cpu' with 'event,period,' that may result in non-fatal errors
> > > inst_retired.any -> cpu/inst_retired.any/
> > > ...
> > > ```
> > >
> > > After:
> > > ```
> > > $ sudo perf stat -vv -e inst_retired.any true
> > > Using CPUID GenuineIntel-6-8D-1
> > > intel_pt default config: tsc,mtc,mtc_period=3,psb_period=3,pt,branch
> > > Attempt to add: cpu/inst_retired.any/
> > > ..after resolving event: cpu/event=0xc0,period=0x1e8483/
> > > inst_retired.any -> cpu/event=0xc0,period=0x1e8483/
> > > ...
> > > ```
> > >
> > > Signed-off-by: Ian Rogers <irogers@xxxxxxxxxx>
> > > ---
> > > tools/perf/util/parse-events.c | 101 ++++++++++++++++++++++++---------
> > > tools/perf/util/parse-events.h | 2 +
> > > tools/perf/util/pmu.c | 19 ++-----
> > > 3 files changed, 81 insertions(+), 41 deletions(-)
> >
> > Which branch should I use to apply the patch?
> > I tried the perf-tools-next branch, but there is some conflict.
> > I'd like to do some tests on a hybrid machine.
>
> I was working on tmp.perf-tools-next but it is the same as
> perf-tools-next currently [1]. I had this 2 line patch in place:
> https://lore.kernel.org/lkml/20230830000545.1638964-1-irogers@xxxxxxxxxx/
> so that may be the conflict.
>
> Thanks,
> Ian
>
> [1] https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/log/?h=tmp.perf-tools-next

Hmm.. for some reason I'm not seeing the message on LKML and so b4
fails. Anyway, applying the patches and running on an Alderlake I see:
```
$ perf stat -vv -e inst_retired.any true
Using CPUID GenuineIntel-6-97-2
Attempt to add: cpu_atom/inst_retired.any/
..after resolving event: cpu_atom/event=0xc0,period=0x1e8483/
inst_retired.any -> cpu_atom/event=0xc0,period=0x1e8483/
Attempt to add: cpu_core/inst_retired.any/
..after resolving event: cpu_core/event=0xc0,period=0x1e8483/
inst_retired.any -> cpu_core/event=0xc0,period=0x1e8483/
...
```
The previous output was like:
```
$ perf stat -vv -e inst_retired.any true
Using CPUID GenuineIntel-6-97-2
Attempting to add event pmu 'cpu_core' with 'inst_retired.any,' that
may result in non-fatal errors
After aliases, add event pmu 'cpu_core' with 'event,period,' that may
result in non-fatal errors
inst_retired.any -> cpu_core/event=0xc0,period=0x1e8483/
Attempting to add event pmu 'cpu_atom' with 'inst_retired.any,' that
may result in non-fatal errors
After aliases, add event pmu 'cpu_atom' with 'event,period,' that may
result in non-fatal errors
inst_retired.any -> cpu_atom/event=0xc0,period=0x1e8483/
...
```
Perhaps a more interesting example is:
```
$ perf stat -vv -e UOPS_RETIRED.MS true
Using CPUID GenuineIntel-6-97-2
Attempt to add: cpu_atom/UOPS_RETIRED.MS/
..after resolving event: cpu_atom/event=0xc2,period=0x1e8483,umask/
UOPS_RETIRED.MS -> cpu_atom/event=0xc2,period=0x1e8483,umask/
Attempt to add: cpu_core/UOPS_RETIRED.MS/
..after resolving event:
cpu_core/event=0xc2,period=0x1e8483,umask=0x4,frontend=0x8/
UOPS_RETIRED.MS -> cpu_core/event=0xc2,period=0x1e8483,umask=0x4,frontend=0x8/
...
```
The umask not being printed for the cpu_atom is an issue. The problem
is how we encode terms of an event name, it is indistinguishable when
the of the user field is 1. I'll probably add something to fix this
later, but it only impacts debug output and perf list, so I'm not
overly worried. We could go in the other direction and instead of
printing say cpu_atom/UOPS_RETIRED.MS/ print
cpu_atom/UOPS_RETIRED.MS=1/, but I thought that was uglier and more
confusing.

Thanks,
Ian

> > Thanks,
> > Kan
> > >
> > > diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
> > > index 4c812fbe0cf9..0b941b58bdc0 100644
> > > --- a/tools/perf/util/parse-events.c
> > > +++ b/tools/perf/util/parse-events.c
> > > @@ -13,7 +13,7 @@
> > > #include <subcmd/parse-options.h>
> > > #include "parse-events.h"
> > > #include "string2.h"
> > > -#include "strlist.h"
> > > +#include "strbuf.h"
> > > #include "debug.h"
> > > #include <api/fs/tracing_path.h>
> > > #include <perf/cpumap.h>
> > > @@ -1303,19 +1303,6 @@ int parse_events_add_pmu(struct parse_events_state *parse_state,
> > >
> > > pmu = parse_state->fake_pmu ?: perf_pmus__find(name);
> > >
> > > - if (verbose > 1 && !(pmu && pmu->selectable)) {
> > > - fprintf(stderr, "Attempting to add event pmu '%s' with '",
> > > - name);
> > > - if (head_config) {
> > > - struct parse_events_term *term;
> > > -
> > > - list_for_each_entry(term, head_config, list) {
> > > - fprintf(stderr, "%s,", term->config);
> > > - }
> > > - }
> > > - fprintf(stderr, "' that may result in non-fatal errors\n");
> > > - }
> > > -
> > > if (!pmu) {
> > > char *err_str;
> > >
> > > @@ -1325,6 +1312,21 @@ int parse_events_add_pmu(struct parse_events_state *parse_state,
> > > parse_events_error__handle(err, loc->first_column, err_str, NULL);
> > > return -EINVAL;
> > > }
> > > +
> > > + if (verbose > 1) {
> > > + struct strbuf sb;
> > > +
> > > + strbuf_init(&sb, /*hint=*/ 0);
> > > + if (pmu->selectable && !head_config) {
> > > + strbuf_addf(&sb, "%s//", name);
> > > + } else {
> > > + strbuf_addf(&sb, "%s/", name);
> > > + parse_events_term__to_strbuf(head_config, &sb);
> > > + strbuf_addch(&sb, '/');
> > > + }
> > > + fprintf(stderr, "Attempt to add: %s\n", sb.buf);
> > > + strbuf_release(&sb);
> > > + }
> > > if (head_config)
> > > fix_raw(head_config, pmu);
> > >
> > > @@ -1349,16 +1351,12 @@ int parse_events_add_pmu(struct parse_events_state *parse_state,
> > > return -EINVAL;
> > >
> > > if (verbose > 1) {
> > > - fprintf(stderr, "After aliases, add event pmu '%s' with '",
> > > - name);
> > > - if (head_config) {
> > > - struct parse_events_term *term;
> > > + struct strbuf sb;
> > >
> > > - list_for_each_entry(term, head_config, list) {
> > > - fprintf(stderr, "%s,", term->config);
> > > - }
> > > - }
> > > - fprintf(stderr, "' that may result in non-fatal errors\n");
> > > + strbuf_init(&sb, /*hint=*/ 0);
> > > + parse_events_term__to_strbuf(head_config, &sb);
> > > + fprintf(stderr, "..after resolving event: %s/%s/\n", name, sb.buf);
> > > + strbuf_release(&sb);
> > > }
> > >
> > > /*
> > > @@ -1460,7 +1458,12 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state,
> > > parse_events_copy_term_list(head, &orig_head);
> > > if (!parse_events_add_pmu(parse_state, list, pmu->name,
> > > orig_head, auto_merge_stats, loc)) {
> > > - pr_debug("%s -> %s/%s/\n", str, pmu->name, str);
> > > + struct strbuf sb;
> > > +
> > > + strbuf_init(&sb, /*hint=*/ 0);
> > > + parse_events_term__to_strbuf(orig_head, &sb);
> > > + pr_debug("%s -> %s/%s/\n", str, pmu->name, sb.buf);
> > > + strbuf_release(&sb);
> > > ok++;
> > > }
> > > parse_events_terms__delete(orig_head);
> > > @@ -1469,7 +1472,12 @@ int parse_events_multi_pmu_add(struct parse_events_state *parse_state,
> > > if (parse_state->fake_pmu) {
> > > if (!parse_events_add_pmu(parse_state, list, str, head,
> > > /*auto_merge_stats=*/true, loc)) {
> > > - pr_debug("%s -> %s/%s/\n", str, "fake_pmu", str);
> > > + struct strbuf sb;
> > > +
> > > + strbuf_init(&sb, /*hint=*/ 0);
> > > + parse_events_term__to_strbuf(head, &sb);
> > > + pr_debug("%s -> %s/%s/\n", str, "fake_pmu", sb.buf);
> > > + strbuf_release(&sb);
> > > ok++;
> > > }
> > > }
> > > @@ -2085,7 +2093,7 @@ void parse_events_error__handle(struct parse_events_error *err, int idx,
> > > break;
> > > default:
> > > pr_debug("Multiple errors dropping message: %s (%s)\n",
> > > - err->str, err->help);
> > > + err->str, err->help ?: "<no help>");
> > > free(err->str);
> > > err->str = str;
> > > free(err->help);
> > > @@ -2502,6 +2510,47 @@ void parse_events_terms__delete(struct list_head *terms)
> > > free(terms);
> > > }
> > >
> > > +int parse_events_term__to_strbuf(struct list_head *term_list, struct strbuf *sb)
> > > +{
> > > + struct parse_events_term *term;
> > > + bool first = true;
> > > +
> > > + if (!term_list)
> > > + return 0;
> > > +
> > > + list_for_each_entry(term, term_list, list) {
> > > + int ret;
> > > +
> > > + if (!first) {
> > > + ret = strbuf_addch(sb, ',');
> > > + if (ret < 0)
> > > + return ret;
> > > + }
> > > + first = false;
> > > +
> > > + if (term->type_val == PARSE_EVENTS__TERM_TYPE_NUM)
> > > + if (term->type_term == PARSE_EVENTS__TERM_TYPE_USER && term->val.num == 1)
> > > + ret = strbuf_addf(sb, "%s", term->config);
> > > + else
> > > + ret = strbuf_addf(sb, "%s=%#"PRIx64, term->config, term->val.num);
> > > + else if (term->type_val == PARSE_EVENTS__TERM_TYPE_STR) {
> > > + if (term->config) {
> > > + ret = strbuf_addf(sb, "%s=", term->config);
> > > + if (ret < 0)
> > > + return ret;
> > > + } else if (term->type_term < __PARSE_EVENTS__TERM_TYPE_NR) {
> > > + ret = strbuf_addf(sb, "%s=", config_term_names[term->type_term]);
> > > + if (ret < 0)
> > > + return ret;
> > > + }
> > > + ret = strbuf_addf(sb, "%s", term->val.str);
> > > + }
> > > + if (ret < 0)
> > > + return ret;
> > > + }
> > > + return 0;
> > > +}
> > > +
> > > void parse_events_evlist_error(struct parse_events_state *parse_state,
> > > int idx, const char *str)
> > > {
> > > diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h
> > > index 6d75d853ce00..20bdc35d6112 100644
> > > --- a/tools/perf/util/parse-events.h
> > > +++ b/tools/perf/util/parse-events.h
> > > @@ -18,6 +18,7 @@ struct parse_events_error;
> > >
> > > struct option;
> > > struct perf_pmu;
> > > +struct strbuf;
> > >
> > > const char *event_type(int type);
> > >
> > > @@ -152,6 +153,7 @@ int parse_events_term__clone(struct parse_events_term **new,
> > > void parse_events_term__delete(struct parse_events_term *term);
> > > void parse_events_terms__delete(struct list_head *terms);
> > > void parse_events_terms__purge(struct list_head *terms);
> > > +int parse_events_term__to_strbuf(struct list_head *term_list, struct strbuf *sb);
> > > int parse_events__modifier_event(struct list_head *list, char *str, bool add);
> > > int parse_events__modifier_group(struct list_head *list, char *event_mod);
> > > int parse_events_name(struct list_head *list, const char *name);
> > > diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c
> > > index b3f8f3f1e900..8dbb7008877e 100644
> > > --- a/tools/perf/util/pmu.c
> > > +++ b/tools/perf/util/pmu.c
> > > @@ -507,12 +507,11 @@ static int perf_pmu__new_alias(struct perf_pmu *pmu, const char *name,
> > > const char *desc, const char *val, FILE *val_fd,
> > > const struct pmu_event *pe)
> > > {
> > > - struct parse_events_term *term;
> > > struct perf_pmu_alias *alias;
> > > int ret;
> > > - char newval[256];
> > > const char *long_desc = NULL, *topic = NULL, *unit = NULL, *pmu_name = NULL;
> > > bool deprecated = false, perpkg = false;
> > > + struct strbuf sb;
> > >
> > > if (perf_pmu__find_alias(pmu, name, /*load=*/ false)) {
> > > /* Alias was already created/loaded. */
> > > @@ -582,20 +581,10 @@ static int perf_pmu__new_alias(struct perf_pmu *pmu, const char *name,
> > > *
> > > * Rebuild string to make alias->str member comparable.
> > > */
> > > - ret = 0;
> > > - list_for_each_entry(term, &alias->terms, list) {
> > > - if (ret)
> > > - ret += scnprintf(newval + ret, sizeof(newval) - ret,
> > > - ",");
> > > - if (term->type_val == PARSE_EVENTS__TERM_TYPE_NUM)
> > > - ret += scnprintf(newval + ret, sizeof(newval) - ret,
> > > - "%s=%#x", term->config, term->val.num);
> > > - else if (term->type_val == PARSE_EVENTS__TERM_TYPE_STR)
> > > - ret += scnprintf(newval + ret, sizeof(newval) - ret,
> > > - "%s=%s", term->config, term->val.str);
> > > - }
> > > zfree(&alias->str);
> > > - alias->str = strdup(newval);
> > > + strbuf_init(&sb, /*hint=*/ 0);
> > > + parse_events_term__to_strbuf(&alias->terms, &sb);
> > > + alias->str = strbuf_detach(&sb, /*sz=*/ NULL);
> > > if (!pe)
> > > pmu->sysfs_aliases++;
> > > else

Next message: Hans Verkuil: "Re: [PATCH v5 02/10] media: videobuf2: Access vb2_queue bufs array through helper functions"
Previous message: Marco Elver: "Re: [PATCH 11/15] stackdepot: use read/write lock"
In reply to: Ian Rogers: "Re: [PATCH v1 2/3] perf parse-events: Make common term list to strbuf helper"
Next in thread: Liang, Kan: "Re: [PATCH v1 2/3] perf parse-events: Make common term list to strbuf helper"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]