| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Thu, 17 Nov 2022 13:31:53 -0300
From: Arnaldo Carvalho de Melo <acme@...nel.org>
To: Ian Rogers <irogers@...gle.com>
Cc: Weilin Wang <weilin.wang@...el.com>,
Perry Taylor <perry.taylor@...el.com>,
Caleb Biggers <caleb.biggers@...el.com>,
Leo Yan <leo.yan@...aro.org>,
Adrian Hunter <adrian.hunter@...el.com>,
Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>,
Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Jiri Olsa <jolsa@...nel.org>,
Namhyung Kim <namhyung@...nel.org>,
Sandipan Das <sandipan.das@....com>,
Kajol Jain <kjain@...ux.ibm.com>,
Zhengjun Xing <zhengjun.xing@...ux.intel.com>,
Kan Liang <kan.liang@...ux.intel.com>,
Ravi Bangoria <ravi.bangoria@....com>,
Xin Gao <gaoxin@...rlc.com>, Rob Herring <robh@...nel.org>,
linux-kernel@...r.kernel.org, linux-perf-users@...r.kernel.org,
Stephane Eranian <eranian@...gle.com>
Subject: Re: [PATCH v3 10/10] perf list: Add json output option
Em Wed, Nov 16, 2022 at 11:52:39AM -0800, Ian Rogers escreveu:
> On Wed, Nov 16, 2022 at 7:21 AM Arnaldo Carvalho de Melo
> <acme@...nel.org> wrote:
> >
> > Em Wed, Nov 16, 2022 at 10:10:53AM -0300, Arnaldo Carvalho de Melo escreveu:
> > > Em Wed, Nov 16, 2022 at 09:41:43AM -0300, Arnaldo Carvalho de Melo escreveu:
> > > > But then:
> > >
> > > > [root@...e ~]# perf list syscalls:sys_enter_open* |& grep syscalls:
> > > > syscalls:sys_enter_open [Tracepoint event]
> > > > syscalls:sys_enter_open_by_handle_at [Tracepoint event]
> > > > syscalls:sys_enter_open_tree [Tracepoint event]
> > > > syscalls:sys_enter_openat [Tracepoint event]
> > > > syscalls:sys_enter_openat2 [Tracepoint event]
> > > > [root@...e ~]#
> > > >
> > > > This stops working, looking into it.
> > >
> > > Sidetracked with other stuff, please find what I have patched at
> > > perf/perf-list-json-output in my tree.
> > >
> > > I removed the last two patches and I'm testing so that I can push
> > > perf/core with your series modulo the last two + Namhyung's 'perf list'
> > > kit.
> >
> > I just saw you sent a patch on top of the previous one, will try and
> > combine stuff to remove failures from the bisect history.
> >
> > - Arnaldo
>
> So the failing test was skipping for me due to a lack of kernel
> symbols, sorry for not spotting it. I find that the issue is resolved
> with your fixes and:
>
> ```
> diff --git a/tools/perf/builtin-list.c b/tools/perf/builtin-list.c
> index 30937e1dd82c..ad6cb5d2e1cc 100644
> --- a/tools/perf/builtin-list.c
> +++ b/tools/perf/builtin-list.c
> @@ -107,7 +107,7 @@ static void default_print_event(void *ps, const
> char *pmu_name, const char *topi
> if (deprecated && !print_state->deprecated)
> return;
>
> - if (print_state->pmu_glob && !strglobmatch(pmu_name,
> print_state->pmu_glob))
> + if (print_state->pmu_glob && pmu_name &&
> !strglobmatch(pmu_name, print_state->pmu_glob))
> return;
>
> if (print_state->event_glob &&
> @@ -534,24 +534,18 @@ int cmd_list(int argc, const char **argv)
> default_ps.metrics = false;
> metricgroup__print(&print_cb, ps);
> } else if ((sep = strchr(argv[i], ':')) != NULL) {
> - int sep_idx;
> -
> - sep_idx = sep - argv[i];
> - s = strdup(argv[i]);
> - if (s == NULL) {
> + default_ps.event_glob = strdup(argv[i]);
> + if (!default_ps.event_glob) {
> ret = -1;
> goto out;
> }
> -
> - s[sep_idx] = '\0';
> - default_ps.pmu_glob = s;
> - default_ps.event_glob = s + sep_idx + 1;
> print_tracepoint_events(&print_cb, ps);
> print_sdt_events(&print_cb, ps);
> default_ps.metrics = true;
> default_ps.metricgroups = true;
> metricgroup__print(&print_cb, ps);
> - free(s);
> + free(default_ps.event_glob);
> + default_ps.event_glob = NULL;
> ```
> I think this should be squashed into "perf list: Reorganize to use
> callbacks". Some explanation, in porting the : glob case I'd assumed
> the before the colon would be the PMU and the after the event. Doing
> things caused tracepoint output to differ too much and so for
> tracepoints the : is kept in the event name. So we can simplify the
> matching to not be pmu and event, just use the event glob.
Next time please send the patch, I did it manually and before the last
option I get:
[root@...co ~]# perf list syscalls:sys_enter_open |& grep syscalls
syscalls:sys_enter_open [Tracepoint event]
[root@...co ~]# perf test 112
112: Check open filename arg using perf trace + vfs_getname : Ok
[root@...co ~]#
diff --git a/tools/perf/builtin-list.c b/tools/perf/builtin-list.c
index 0c84fdb3ad37c1ad..7b63bc30665a7f56 100644
--- a/tools/perf/builtin-list.c
+++ b/tools/perf/builtin-list.c
@@ -19,6 +19,7 @@
#include "util/strlist.h"
#include <subcmd/pager.h>
#include <subcmd/parse-options.h>
+#include <linux/zalloc.h>
#include <stdio.h>
/**
@@ -328,25 +329,20 @@ int cmd_list(int argc, const char **argv)
ps.metrics = false;
metricgroup__print(&print_cb, &ps);
} else if ((sep = strchr(argv[i], ':')) != NULL) {
- int sep_idx;
char *old_pmu_glob = ps.pmu_glob;
- sep_idx = sep - argv[i];
- s = strdup(argv[i]);
- if (s == NULL) {
+ ps.event_glob = strdup(argv[i]);
+ if (!ps.event_glob) {
ret = -1;
goto out;
}
- s[sep_idx] = '\0';
- ps.pmu_glob = s;
- ps.event_glob = s + sep_idx + 1;
print_tracepoint_events(&print_cb, &ps);
print_sdt_events(&print_cb, &ps);
ps.metrics = true;
ps.metricgroups = true;
metricgroup__print(&print_cb, &ps);
- free(s);
+ zfree(&ps.event_glob);
ps.pmu_glob = old_pmu_glob;
} else {
if (asprintf(&s, "*%s*", argv[i]) < 0) {
Powered by blists - more mailing lists