| 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
| ||
|
Message-Id: <1332950153-8606-5-git-send-email-jolsa@redhat.com>
Date: Wed, 28 Mar 2012 17:55:52 +0200
From: Jiri Olsa <jolsa@...hat.com>
To: acme@...hat.com, a.p.zijlstra@...llo.nl, mingo@...e.hu,
paulus@...ba.org, cjashfor@...ux.vnet.ibm.com, fweisbec@...il.com
Cc: linux-kernel@...r.kernel.org, Jiri Olsa <jolsa@...hat.com>
Subject: [PATCH 4/5] perf, tool: Add pmu event parse support for branch_sample_type values
Adding a support to specify branch_type as a hardcoded term
inside the pmu event definition.
It is possible to specify pmu event like:
"cpu/config=1,branch_type=hv|any_ret|1,config2=2/u"
Following string values could be used as value for branch_type:
u (PERF_SAMPLE_BRANCH_USER)
k (PERF_SAMPLE_BRANCH_KERNEL)
hv (PERF_SAMPLE_BRANCH_HV)
any (PERF_SAMPLE_BRANCH_ANY)
any_call (PERF_SAMPLE_BRANCH_ANY_CALL)
any_ret (PERF_SAMPLE_BRANCH_ANY_RETURN)
ind_call (PERF_SAMPLE_BRANCH_IND_CALL)
Also a number could be specified as value.
Signed-off-by: Jiri Olsa <jolsa@...hat.com>
---
tools/perf/builtin-record.c | 31 +-----------
tools/perf/builtin-test.c | 27 +++++++++++
tools/perf/util/parse-events.c | 100 ++++++++++++++++++++++++++++++++++++++--
tools/perf/util/parse-events.h | 1 +
4 files changed, 127 insertions(+), 32 deletions(-)
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index be4e1ee..87fc68d 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -641,27 +641,6 @@ out_delete_session:
return err;
}
-#define BRANCH_OPT(n, m) \
- { .name = n, .mode = (m) }
-
-#define BRANCH_END { .name = NULL }
-
-struct branch_mode {
- const char *name;
- int mode;
-};
-
-static const struct branch_mode branch_modes[] = {
- BRANCH_OPT("u", PERF_SAMPLE_BRANCH_USER),
- BRANCH_OPT("k", PERF_SAMPLE_BRANCH_KERNEL),
- BRANCH_OPT("hv", PERF_SAMPLE_BRANCH_HV),
- BRANCH_OPT("any", PERF_SAMPLE_BRANCH_ANY),
- BRANCH_OPT("any_call", PERF_SAMPLE_BRANCH_ANY_CALL),
- BRANCH_OPT("any_ret", PERF_SAMPLE_BRANCH_ANY_RETURN),
- BRANCH_OPT("ind_call", PERF_SAMPLE_BRANCH_IND_CALL),
- BRANCH_END
-};
-
static int
parse_branch_stack(const struct option *opt, const char *str, int unset)
{
@@ -671,7 +650,6 @@ parse_branch_stack(const struct option *opt, const char *str, int unset)
PERF_SAMPLE_BRANCH_HV)
uint64_t *mode = (uint64_t *)opt->value;
- const struct branch_mode *br;
char *s, *os = NULL, *p;
int ret = -1;
@@ -692,21 +670,18 @@ parse_branch_stack(const struct option *opt, const char *str, int unset)
return -1;
for (;;) {
+ u64 bmode;
p = strchr(s, ',');
if (p)
*p = '\0';
- for (br = branch_modes; br->name; br++) {
- if (!strcasecmp(s, br->name))
- break;
- }
- if (!br->name) {
+ if (parse_events__branch_type(&bmode, s)) {
ui__warning("unknown branch filter %s,"
" check man page\n", s);
goto error;
}
- *mode |= br->mode;
+ *mode |= bmode;
if (!p)
break;
diff --git a/tools/perf/builtin-test.c b/tools/perf/builtin-test.c
index d522cf4..2e4dc0a 100644
--- a/tools/perf/builtin-test.c
+++ b/tools/perf/builtin-test.c
@@ -1008,6 +1008,29 @@ static int test__checkevent_pmu_name(struct perf_evlist *evlist)
return 0;
}
+static int test__checkevent_pmu_branch_type(struct perf_evlist *evlist)
+{
+ struct perf_evsel *evsel = list_entry(evlist->entries.next,
+ struct perf_evsel, node);
+
+ TEST_ASSERT_VAL("wrong number of entries", 1 == evlist->nr_entries);
+ TEST_ASSERT_VAL("wrong type", PERF_TYPE_RAW == evsel->attr.type);
+ TEST_ASSERT_VAL("wrong config", 1 == evsel->attr.config);
+ TEST_ASSERT_VAL("wrong config2", 2 == evsel->attr.config2);
+
+#define BT (PERF_SAMPLE_BRANCH_USER | \
+ PERF_SAMPLE_BRANCH_HV | \
+ PERF_SAMPLE_BRANCH_ANY_RETURN)
+
+ TEST_ASSERT_VAL("wrong sample_type",
+ PERF_SAMPLE_BRANCH_STACK && evsel->attr.sample_type);
+ TEST_ASSERT_VAL("wrong branch_type",
+ BT == evsel->attr.branch_sample_type);
+
+#undef BT
+ return 0;
+}
+
static struct test__event_st {
const char *name;
__u32 type;
@@ -1117,6 +1140,10 @@ static struct test__event_st {
.name = "cpu/config=1,name=krava/u,cpu/config=2/u",
.check = test__checkevent_pmu_name,
},
+ {
+ .name = "cpu/config=1,branch_type=hv|any_ret|1,config2=2/u",
+ .check = test__checkevent_pmu_branch_type,
+ },
};
#define TEST__EVENTS_CNT (sizeof(test__events) / sizeof(struct test__event_st))
diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
index 29892c0..3e99637 100644
--- a/tools/perf/util/parse-events.c
+++ b/tools/perf/util/parse-events.c
@@ -590,6 +590,99 @@ int parse_events_add_breakpoint(struct list_head *list, int *idx,
return add_event(list, idx, &attr, name);
}
+#define BRANCH_OPT(n, m) { .name = n, .mode = (m) }
+#define BRANCH_END { .name = NULL }
+
+struct branch_mode {
+ const char *name;
+ int mode;
+};
+
+static const struct branch_mode branch_modes[] = {
+ BRANCH_OPT("u", PERF_SAMPLE_BRANCH_USER),
+ BRANCH_OPT("k", PERF_SAMPLE_BRANCH_KERNEL),
+ BRANCH_OPT("hv", PERF_SAMPLE_BRANCH_HV),
+ BRANCH_OPT("any", PERF_SAMPLE_BRANCH_ANY),
+ BRANCH_OPT("any_call", PERF_SAMPLE_BRANCH_ANY_CALL),
+ BRANCH_OPT("any_ret", PERF_SAMPLE_BRANCH_ANY_RETURN),
+ BRANCH_OPT("ind_call", PERF_SAMPLE_BRANCH_IND_CALL),
+ BRANCH_END
+};
+
+int parse_events__branch_type(u64 *_type, char *str)
+{
+ const struct branch_mode *br;
+
+ for (br = branch_modes; br->name; br++) {
+ if (!strcasecmp(str, br->name)) {
+ *_type = br->mode;
+ return 0;
+ }
+ }
+
+ return -EINVAL;
+}
+
+typedef int (process_term_list_cb)(struct parse_events__term_elem *elem,
+ void *arg);
+
+static int branch_type_cb(struct parse_events__term_elem *elem, void *arg)
+{
+ u64 *_type = arg;
+ u64 type;
+
+ switch (elem->type_val) {
+ case PARSE_EVENTS__TERM_TYPE_NUM:
+ type = elem->val.num;
+ break;
+ case PARSE_EVENTS__TERM_TYPE_STR:
+ if (parse_events__branch_type(&type, elem->val.str))
+ return -EINVAL;
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ *_type |= type;
+ return 0;
+}
+
+static int process_term_list(struct parse_events__term *term,
+ process_term_list_cb cb, void *arg)
+{
+ struct parse_events__term_elem *elem;
+
+ list_for_each_entry(elem, &term->val.list, list)
+ if (cb(elem, arg))
+ return -EINVAL;
+
+ return 0;
+}
+
+static int branch_type(struct parse_events__term *term, u64 *_type)
+{
+ u64 type = 0;
+
+ switch (term->type_val) {
+ case PARSE_EVENTS__TERM_TYPE_NUM:
+ type = term->val.num;
+ break;
+ case PARSE_EVENTS__TERM_TYPE_STR:
+ if (parse_events__branch_type(&type, term->val.str))
+ return -EINVAL;
+ break;
+ case PARSE_EVENTS__TERM_TYPE_LIST:
+ if (process_term_list(term, branch_type_cb, &type))
+ return -EINVAL;
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ *_type = type;
+ return 0;
+}
+
static int config_term(struct perf_event_attr *attr,
struct parse_events__term *term)
{
@@ -617,10 +710,9 @@ do { \
attr->sample_period = term->val.num;
break;
case PARSE_EVENTS__TERM_TYPE_BRANCH_SAMPLE_TYPE:
- /*
- * TODO uncomment when the field is available
- * attr->branch_sample_type = term->val.num;
- */
+ attr->sample_type |= PERF_SAMPLE_BRANCH_STACK;
+ if (branch_type(term, (u64 *)&attr->branch_sample_type))
+ return -EINVAL;
break;
default:
return -EINVAL;
diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h
index cabcbc1..8bb673d 100644
--- a/tools/perf/util/parse-events.h
+++ b/tools/perf/util/parse-events.h
@@ -83,6 +83,7 @@ int parse_events__term_list(struct parse_events__term **_term,
struct list_head *list);
int parse_events__term_elem(struct parse_events__term_elem **_e,
int type_val, char *str, long num);
+int parse_events__branch_type(u64 *_type, char *str);
void parse_events__free_terms(struct list_head *terms);
int parse_events_modifier(struct list_head *list __used, char *str __used);
int parse_events_add_tracepoint(struct list_head *list, int *idx,
--
1.7.6.5
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists