[Ltrace-devel] [PATCH 1/6] Fix a memory leak in protolib_cache_maybe_load

[Petr Machata]

Роман Донченко <dpb at corrigendum.ru> writes:

>>> -	if (DICT_FIND_VAL(&cache->protolibs, &key, retp) == 0)
>>> +	if (DICT_FIND_VAL(&cache->protolibs, &key, retp) == 0) {
>>> +		if (*retp != NULL && own_key)
>>> +			free((void *) key);
>>>  		return 0;
>>> +	}
>>>
>>>  	if (strdup_if(&key, key, !own_key) < 0) {
>>>  		fprintf(stderr, "Couldn't cache %s: %s\n",
>>
>> This looks like the wrong fix.  The one who allocated the key should
>> free it if protolib_cache_maybe_load returns a failure.
>
> But it returns a success here.

You are right, I didn't notice.  Then it's fine.  I'll push later today
together with the other fixes.

> In fact, I think we should take ownership of the key whenever the
> function returns zero, no matter what's in *retp. What do you think?

Yeah, if a key is passed in with own_key != 0, and the function returns
zero (i.e. a success), the caller should assume the key has been
consumed.

Thanks,
Petr