[PATCH v2] perf expr: Force encapsulation on expr_id_data

From: Ian Rogers
Date: Wed Aug 26 2020 - 11:31:11 EST


This patch resolves some undefined behavior where variables in
expr_id_data were accessed (for debugging) without being defined. To
better enforce the tagged union behavior, the struct is moved into
expr.c and accessors provided. Tag values (kinds) are explicitly
identified.

Signed-off-by: Ian Rogers <irogers@xxxxxxxxxx>
---
tools/perf/util/expr.c | 68 ++++++++++++++++++++++++++++++-----
tools/perf/util/expr.h | 17 +++------
tools/perf/util/expr.y | 2 +-
tools/perf/util/metricgroup.c | 4 +--
4 files changed, 66 insertions(+), 25 deletions(-)

diff --git a/tools/perf/util/expr.c b/tools/perf/util/expr.c
index 53482ef53c41..a850fd0be3ee 100644
--- a/tools/perf/util/expr.c
+++ b/tools/perf/util/expr.c
@@ -17,6 +17,29 @@
extern int expr_debug;
#endif

+struct expr_id_data {
+ union {
+ double val;
+ struct {
+ double val;
+ const char *metric_name;
+ const char *metric_expr;
+ } ref;
+ struct expr_id *parent;
+ };
+
+ enum {
+ /* Holding a double value. */
+ EXPR_ID_DATA__VALUE,
+ /* Reference to another metric. */
+ EXPR_ID_DATA__REF,
+ /* A reference but the value has been computed. */
+ EXPR_ID_DATA__REF_VALUE,
+ /* A parent is remembered for the recursion check. */
+ EXPR_ID_DATA__PARENT,
+ } kind;
+};
+
static size_t key_hash(const void *key, void *ctx __maybe_unused)
{
const char *str = (const char *)key;
@@ -48,6 +71,7 @@ int expr__add_id(struct expr_parse_ctx *ctx, const char *id)
return -ENOMEM;

data_ptr->parent = ctx->parent;
+ data_ptr->kind = EXPR_ID_DATA__PARENT;

ret = hashmap__set(&ctx->ids, id, data_ptr,
(const void **)&old_key, (void **)&old_data);
@@ -69,7 +93,7 @@ int expr__add_id_val(struct expr_parse_ctx *ctx, const char *id, double val)
if (!data_ptr)
return -ENOMEM;
data_ptr->val = val;
- data_ptr->is_ref = false;
+ data_ptr->kind = EXPR_ID_DATA__VALUE;

ret = hashmap__set(&ctx->ids, id, data_ptr,
(const void **)&old_key, (void **)&old_data);
@@ -114,8 +138,7 @@ int expr__add_ref(struct expr_parse_ctx *ctx, struct metric_ref *ref)
*/
data_ptr->ref.metric_name = ref->metric_name;
data_ptr->ref.metric_expr = ref->metric_expr;
- data_ptr->ref.counted = false;
- data_ptr->is_ref = true;
+ data_ptr->kind = EXPR_ID_DATA__REF;

ret = hashmap__set(&ctx->ids, name, data_ptr,
(const void **)&old_key, (void **)&old_data);
@@ -148,17 +171,30 @@ int expr__resolve_id(struct expr_parse_ctx *ctx, const char *id,

data = *datap;

- pr_debug2("lookup: is_ref %d, counted %d, val %f: %s\n",
- data->is_ref, data->ref.counted, data->val, id);
-
- if (data->is_ref && !data->ref.counted) {
- data->ref.counted = true;
+ switch (data->kind) {
+ case EXPR_ID_DATA__VALUE:
+ pr_debug2("lookup(%s): val %f\n", id, data->val);
+ break;
+ case EXPR_ID_DATA__PARENT:
+ pr_debug2("lookup(%s): parent %s\n", id, data->parent->id);
+ break;
+ case EXPR_ID_DATA__REF:
+ pr_debug2("lookup(%s): ref metric name %s\n", id,
+ data->ref.metric_name);
pr_debug("processing metric: %s ENTRY\n", id);
- if (expr__parse(&data->val, ctx, data->ref.metric_expr, 1)) {
+ data->kind = EXPR_ID_DATA__REF_VALUE;
+ if (expr__parse(&data->ref.val, ctx, data->ref.metric_expr, 1)) {
pr_debug("%s failed to count\n", id);
return -1;
}
pr_debug("processing metric: %s EXIT: %f\n", id, data->val);
+ break;
+ case EXPR_ID_DATA__REF_VALUE:
+ pr_debug2("lookup(%s): ref val %f metric name %s\n", id,
+ data->ref.val, data->ref.metric_name);
+ break;
+ default:
+ assert(0); /* Unreachable. */
}

return 0;
@@ -241,3 +277,17 @@ int expr__find_other(const char *expr, const char *one,

return ret;
}
+
+double expr_id_data__value(const struct expr_id_data *data)
+{
+ if (data->kind == EXPR_ID_DATA__VALUE)
+ return data->val;
+ assert(data->kind == EXPR_ID_DATA__REF_VALUE);
+ return data->ref.val;
+}
+
+struct expr_id *expr_id_data__parent(struct expr_id_data *data)
+{
+ assert(data->kind == EXPR_ID_DATA__PARENT);
+ return data->parent;
+}
diff --git a/tools/perf/util/expr.h b/tools/perf/util/expr.h
index fc2b5e824a66..dcf8d19b83c8 100644
--- a/tools/perf/util/expr.h
+++ b/tools/perf/util/expr.h
@@ -23,19 +23,7 @@ struct expr_parse_ctx {
struct expr_id *parent;
};

-struct expr_id_data {
- union {
- double val;
- struct {
- const char *metric_name;
- const char *metric_expr;
- bool counted;
- } ref;
- struct expr_id *parent;
- };
-
- bool is_ref;
-};
+struct expr_id_data;

struct expr_scanner_ctx {
int start_token;
@@ -57,4 +45,7 @@ int expr__parse(double *final_val, struct expr_parse_ctx *ctx,
int expr__find_other(const char *expr, const char *one,
struct expr_parse_ctx *ids, int runtime);

+double expr_id_data__value(const struct expr_id_data *data);
+struct expr_id *expr_id_data__parent(struct expr_id_data *data);
+
#endif
diff --git a/tools/perf/util/expr.y b/tools/perf/util/expr.y
index d34b370391c6..b2ada8f8309a 100644
--- a/tools/perf/util/expr.y
+++ b/tools/perf/util/expr.y
@@ -93,7 +93,7 @@ expr: NUMBER
YYABORT;
}

- $$ = data->val;
+ $$ = expr_id_data__value(data);
free($1);
}
| expr '|' expr { $$ = (long)$1 | (long)$3; }
diff --git a/tools/perf/util/metricgroup.c b/tools/perf/util/metricgroup.c
index 8831b964288f..339bfb19a10b 100644
--- a/tools/perf/util/metricgroup.c
+++ b/tools/perf/util/metricgroup.c
@@ -786,7 +786,7 @@ static int recursion_check(struct metric *m, const char *id, struct expr_id **pa
if (ret)
return ret;

- p = data->parent;
+ p = expr_id_data__parent(data);

while (p->parent) {
if (!strcmp(p->id, id)) {
@@ -807,7 +807,7 @@ static int recursion_check(struct metric *m, const char *id, struct expr_id **pa
}

p->id = strdup(id);
- p->parent = data->parent;
+ p->parent = expr_id_data__parent(data);
*parent = p;

return p->id ? 0 : -ENOMEM;
--
2.28.0.297.g1956fa8f8d-goog