LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: "Liang, Kan" <kan.liang@linux.intel.com>
To: Jiri Olsa <jolsa@redhat.com>, Andi Kleen <ak@linux.intel.com>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
Adrian Hunter <adrian.hunter@intel.com>
Cc: Ingo Molnar <mingo@kernel.org>,
Clark Williams <williams@redhat.com>,
linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org,
Jiri Olsa <jolsa@kernel.org>,
Alexander Shishkin <alexander.shishkin@linux.intel.com>,
David Ahern <dsahern@gmail.com>,
Namhyung Kim <namhyung@kernel.org>,
Peter Zijlstra <peterz@infradead.org>,
Arnaldo Carvalho de Melo <acme@redhat.com>
Subject: Re: [PATCH] perf tools: Fix parser for empty pmu terms case
Date: Sun, 6 May 2018 12:34:48 -0400 [thread overview]
Message-ID: <6a9f5464-eae8-c22e-d518-215c4c30019f@linux.intel.com> (raw)
In-Reply-To: <20180506142830.GA18865@krava>
On 5/6/2018 10:28 AM, Jiri Olsa wrote:
> On Sat, May 05, 2018 at 08:43:11PM -0700, Andi Kleen wrote:
>> Jiri Olsa <jolsa@redhat.com> writes:
>>
>> Please fix this quickly, PT is currently totally non functional in Linus
>> mainline.
>
> attached.. Kan, could you please test it wrt your latest changes?
>
> thanks,
> jirka
>
>
> ---
> Adrian reported broken event parsing for Intel PT:
>
> $ perf record -e intel_pt//u uname
> event syntax error: 'intel_pt//u'
> \___ parser error
> Run 'perf list' for a list of valid events
>
> It's caused by recent change in parsing grammar
> (see Fixes: for commit).
>
> Adding special rule with empty terms config to handle
> the reported case and moving the common rule code into
> new parse_events_pmu function.
>
> Fixes: 9a4a931ce847 ("perf pmu: Fix pmu events parsing rule")
> Reported-by: Adrian Hunter <adrian.hunter@intel.com>
> Cc: Andi Kleen <ak@linux.intel.com>
> Link: http://lkml.kernel.org/n/tip-uorb0azuem7b7ydace7cf6vc@git.kernel.org
> Signed-off-by: Jiri Olsa <jolsa@kernel.org>
The patch looks good to me.
I tested my latest change and Intel PT. Both looks good.
./perf stat -e '{unc_m_cas_count.all,unc_m_clockticks}' -a -I 1000 #
time counts unit events
1.001556622 95,686 unc_m_cas_count.all
1.001556622 1,900,544,149 unc_m_clockticks
2.002722048 159,598 unc_m_cas_count.all
2.002722048 2,339,028,614 unc_m_clockticks
./perf record -e intel_pt//u
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.044 MB perf.data ]
Tested-by: Kan Liang <kan.liang@linux.intel.com>
> ---
> tools/perf/util/parse-events.c | 46 +++++++++++++++++++++++++++++++++++++
> tools/perf/util/parse-events.h | 3 +++
> tools/perf/util/parse-events.y | 51 +++++++++++++-----------------------------
> 3 files changed, 65 insertions(+), 35 deletions(-)
>
> diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
> index 3576aaaa9e4c..7cf326a8effe 100644
> --- a/tools/perf/util/parse-events.c
> +++ b/tools/perf/util/parse-events.c
> @@ -8,6 +8,7 @@
> #include <sys/stat.h>
> #include <fcntl.h>
> #include <sys/param.h>
> +#include <fnmatch.h>
> #include "term.h"
> #include "../perf.h"
> #include "evlist.h"
> @@ -1294,6 +1295,51 @@ int parse_events_add_pmu(struct parse_events_state *parse_state,
> return evsel ? 0 : -ENOMEM;
> }
>
> +int parse_events_pmu(struct parse_events_state *parse_state,
> + struct list_head *list, char *name,
> + struct list_head *head_config)
> +{
> + struct list_head *orig_terms;
> + struct perf_pmu *pmu = NULL;
> + char *pattern = NULL;
> + int ok = 0;
> +
> + if (!parse_events_add_pmu(parse_state, list, name,
> + head_config, false, false))
> + return 0;
> +
> + if (parse_events_copy_term_list(head_config, &orig_terms))
> + return -1;
> +
> + if (asprintf(&pattern, "%s*", name) < 0)
> + goto out;
> +
> + while ((pmu = perf_pmu__scan(pmu)) != NULL) {
> + char *tmp = pmu->name;
> +
> + if (!strncmp(tmp, "uncore_", 7) &&
> + strncmp(name, "uncore_", 7))
> + tmp += 7;
> + if (!fnmatch(pattern, tmp, 0)) {
> + struct list_head *terms;
> +
> + if (parse_events_copy_term_list(orig_terms, &terms)) {
> + ok = 0;
> + goto out;
> + }
> + if (!parse_events_add_pmu(parse_state, list, pmu->name,
> + terms, true, false))
> + ok++;
> + parse_events_terms__delete(terms);
> + }
> + }
> +
> +out:
> + parse_events_terms__delete(orig_terms);
> + free(pattern);
> + return ok ? 0 : -1;
> +}
> +
> int parse_events_multi_pmu_add(struct parse_events_state *parse_state,
> char *str, struct list_head **listp)
> {
> diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h
> index 4473dac27aee..553fcaf5d23e 100644
> --- a/tools/perf/util/parse-events.h
> +++ b/tools/perf/util/parse-events.h
> @@ -170,6 +170,9 @@ int parse_events_add_pmu(struct parse_events_state *parse_state,
> struct list_head *head_config,
> bool auto_merge_stats,
> bool use_alias);
> +int parse_events_pmu(struct parse_events_state *parse_state,
> + struct list_head *list, char *name,
> + struct list_head *head_config);
>
> int parse_events_multi_pmu_add(struct parse_events_state *parse_state,
> char *str,
> diff --git a/tools/perf/util/parse-events.y b/tools/perf/util/parse-events.y
> index 47f6399a309a..d44599bece43 100644
> --- a/tools/perf/util/parse-events.y
> +++ b/tools/perf/util/parse-events.y
> @@ -8,7 +8,6 @@
>
> #define YYDEBUG 1
>
> -#include <fnmatch.h>
> #include <linux/compiler.h>
> #include <linux/list.h>
> #include <linux/types.h>
> @@ -226,44 +225,26 @@ event_def: event_pmu |
> event_pmu:
> PE_NAME '/' event_config '/'
> {
> - struct list_head *list, *orig_terms, *terms;
> + struct list_head *list;
> +
> + ALLOC_LIST(list);
>
> - if (parse_events_copy_term_list($3, &orig_terms))
> + if (parse_events_pmu(_parse_state, list, $1, $3))
> YYABORT;
>
> - ALLOC_LIST(list);
> - if (parse_events_add_pmu(_parse_state, list, $1, $3, false, false)) {
> - struct perf_pmu *pmu = NULL;
> - int ok = 0;
> - char *pattern;
> -
> - if (asprintf(&pattern, "%s*", $1) < 0)
> - YYABORT;
> -
> - while ((pmu = perf_pmu__scan(pmu)) != NULL) {
> - char *name = pmu->name;
> -
> - if (!strncmp(name, "uncore_", 7) &&
> - strncmp($1, "uncore_", 7))
> - name += 7;
> - if (!fnmatch(pattern, name, 0)) {
> - if (parse_events_copy_term_list(orig_terms, &terms)) {
> - free(pattern);
> - YYABORT;
> - }
> - if (!parse_events_add_pmu(_parse_state, list, pmu->name, terms, true, false))
> - ok++;
> - parse_events_terms__delete(terms);
> - }
> - }
> -
> - free(pattern);
> -
> - if (!ok)
> - YYABORT;
> - }
> parse_events_terms__delete($3);
> - parse_events_terms__delete(orig_terms);
> + $$ = list;
> +}
> +|
> +PE_NAME '/' '/'
> +{
> + struct list_head *list;
> +
> + ALLOC_LIST(list);
> +
> + if (parse_events_pmu(_parse_state, list, $1, NULL))
> + YYABORT;
> +
> $$ = list;
> }
> |
>
next prev parent reply other threads:[~2018-05-06 16:34 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-04-25 15:59 [GIT PULL 00/12] perf/urgent fixes Arnaldo Carvalho de Melo
2018-04-25 15:59 ` [PATCH 01/12] perf machine: Set main kernel end address properly Arnaldo Carvalho de Melo
2018-04-25 15:59 ` [PATCH 02/12] perf list: Remove s390 specific strcmp_cpuid_cmp function Arnaldo Carvalho de Melo
2018-04-25 15:59 ` [PATCH 03/12] perf test: Adapt test case record+probe_libc_inet_pton.sh for s390 Arnaldo Carvalho de Melo
2018-04-25 16:00 ` [PATCH 04/12] perf stat: Keep the / modifier separator in fallback Arnaldo Carvalho de Melo
2018-04-25 16:00 ` [PATCH 05/12] perf pmu: Fix pmu events parsing rule Arnaldo Carvalho de Melo
2018-05-03 8:25 ` Adrian Hunter
2018-05-03 10:37 ` Jiri Olsa
2018-05-03 11:38 ` Adrian Hunter
2018-05-03 11:47 ` Jiri Olsa
2018-05-04 16:02 ` Jiri Olsa
2018-05-06 3:43 ` Andi Kleen
2018-05-06 14:28 ` [PATCH] perf tools: Fix parser for empty pmu terms case Jiri Olsa
2018-05-06 16:34 ` Liang, Kan [this message]
2018-05-07 7:21 ` Adrian Hunter
2018-05-07 8:12 ` Jiri Olsa
2018-05-07 15:04 ` Arnaldo Carvalho de Melo
2018-05-07 18:37 ` [PATCH 05/12] perf pmu: Fix pmu events parsing rule Arnaldo Carvalho de Melo
2018-05-07 19:16 ` Jiri Olsa
2018-05-07 19:26 ` Arnaldo Carvalho de Melo
2018-05-07 19:24 ` Arnaldo Carvalho de Melo
2018-05-07 19:42 ` Jiri Olsa
2018-04-25 16:00 ` [PATCH 06/12] perf evsel: Disable write_backward for leader sampling group events Arnaldo Carvalho de Melo
2018-04-25 16:00 ` [PATCH 07/12] perf mem: Document incorrect and missing options Arnaldo Carvalho de Melo
2018-04-25 16:00 ` [PATCH 08/12] perf record: Fix s390 undefined record__auxtrace_init() return value Arnaldo Carvalho de Melo
2018-04-25 16:00 ` [PATCH 09/12] perf pmu: Fix core PMU alias list for X86 platform Arnaldo Carvalho de Melo
2018-04-25 16:00 ` [PATCH 10/12] perf stat: Print out hint for mixed PMU group error Arnaldo Carvalho de Melo
2018-04-25 16:00 ` [PATCH 11/12] perf evsel: Only fall back group read for leader Arnaldo Carvalho de Melo
2018-04-25 16:00 ` [PATCH 12/12] perf stat: Fix duplicate PMU name for interval print Arnaldo Carvalho de Melo
2018-04-26 5:33 ` [GIT PULL 00/12] perf/urgent fixes Ingo Molnar
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=6a9f5464-eae8-c22e-d518-215c4c30019f@linux.intel.com \
--to=kan.liang@linux.intel.com \
--cc=acme@kernel.org \
--cc=acme@redhat.com \
--cc=adrian.hunter@intel.com \
--cc=ak@linux.intel.com \
--cc=alexander.shishkin@linux.intel.com \
--cc=dsahern@gmail.com \
--cc=jolsa@kernel.org \
--cc=jolsa@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=namhyung@kernel.org \
--cc=peterz@infradead.org \
--cc=williams@redhat.com \
--subject='Re: [PATCH] perf tools: Fix parser for empty pmu terms case' \
/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
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).