On 12/02/2014 08:53 PM, Mike Holmes wrote:
On 1 December 2014 at 12:05, Taras Kondratiuk
<[email protected] <mailto:[email protected]>> wrote:
    diff --git a/test/validation/common/odp_cunit_common.h
    b/test/validation/common/odp_cunit_common.h
    index 1f30788..f967ca2 100644
    --- a/test/validation/common/odp_cunit_common.h
    +++ b/test/validation/common/odp_cunit_common.h
    @@ -40,5 +40,19 @@ typedef struct {
      /** create thread fro start_routine function */
      extern int odp_cunit_thread_create(void *func_ptr(void *),
    pthrd_arg *arg);
      extern int odp_cunit_thread_exit(pthrd_arg *);
    +/**
    + * Global tests initialization.
    + *
    + * Initialize global resources needed by all testsuites. Default
    weak definition
    + * do nothing. Test application can override it by defining a
    strong version.
    + * The function is called by the common main() just after ODP
    global/local
    + * initialization.
    + *
    + * @note: This function is a workaround for Crypto test and other
    applications
    + *        should try not to use it, because it will complicate
    migration to a
    + *        single test application in future. Normally each
    testsuite have to
    + *        prepare its environment in its own init function.
    + */
    +extern int tests_global_init(void);


Is this not a case where cryptos odp_testsuite would be altered to
provide this function cryptos CUnit test suite init ?

Crypto should also provide a suite terminate that is symmetrical

CU_SuiteInfo odp_testsuites[] = {
         {ODP_CRYPTO_SYNC_INP, cryptotests_global_init,
cryptotests_global_term, NULL, NULL, test_array_sync },
         {ODP_CRYPTO_ASYNC_INP, cryptotests_global_init,
cryptotests_global_term , NULL, NULL, test_array_async },
         CU_SUITE_INFO_NULL,
};

Exactly, but we don't have necessary API in place to correctly cleanup
resources initialized in testsuite init function. We need at least
buffer and queue destroy API. Patches for them are in ML.

_______________________________________________
lng-odp mailing list
[email protected]
http://lists.linaro.org/mailman/listinfo/lng-odp

Reply via email to