From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:60589) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Ze96R-0007ro-Gb for qemu-devel@nongnu.org; Mon, 21 Sep 2015 17:58:28 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Ze96P-0000ze-UO for qemu-devel@nongnu.org; Mon, 21 Sep 2015 17:58:27 -0400 Received: from mx1.redhat.com ([209.132.183.28]:48484) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Ze96P-0000zR-P9 for qemu-devel@nongnu.org; Mon, 21 Sep 2015 17:58:25 -0400 From: Eric Blake Date: Mon, 21 Sep 2015 15:57:42 -0600 Message-Id: <1442872682-6523-27-git-send-email-eblake@redhat.com> In-Reply-To: <1442872682-6523-1-git-send-email-eblake@redhat.com> References: <1442872682-6523-1-git-send-email-eblake@redhat.com> Subject: [Qemu-devel] [PATCH v5 26/46] qapi: Test failure in middle of array parse List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: Michael Roth , marcandre.lureau@redhat.com, DirtY.iCE.hu@gmail.com, armbru@redhat.com, ehabkost@redhat.com Our generated list visitors have the same problem as has been mentioned elsewhere (see commit 2f52e20): they allocate data even on failure. An upcoming patch will correct things to provide saner guarantees, but first we need to expose the behavior in the testsuite to ensure we aren't introducing any memory usage bugs. Signed-off-by: Eric Blake --- scripts/qapi-visit.py | 4 ++++ tests/test-qmp-input-visitor.c | 13 +++++++++++++ 2 files changed, 17 insertions(+) diff --git a/scripts/qapi-visit.py b/scripts/qapi-visit.py index 6ca7c7c..55f6430 100644 --- a/scripts/qapi-visit.py +++ b/scripts/qapi-visit.py @@ -132,6 +132,10 @@ void visit_type_%(c_name)s(Visitor *v, %(c_name)s **obj, const char *name, Error def gen_visit_list(name, element_type): + # FIXME: if *obj is NULL on entry, and the first visit_next_list() + # assigns to *obj, while a later one fails, we should clean up *obj + # rather than leaving it non-NULL. As currently written, the caller must + # call qapi_free_FOOList() to avoid a memory leak of the partial FOOList. return mcgen(''' void visit_type_%(c_name)s(Visitor *v, %(c_name)s **obj, const char *name, Error **errp) diff --git a/tests/test-qmp-input-visitor.c b/tests/test-qmp-input-visitor.c index 0c236f0..667301c 100644 --- a/tests/test-qmp-input-visitor.c +++ b/tests/test-qmp-input-visitor.c @@ -819,6 +819,7 @@ static void test_visitor_in_errors(TestInputVisitorData *data, TestStruct *p = NULL; Error *err = NULL; Visitor *v; + strList *q = NULL; v = visitor_input_test_init(data, "{ 'integer': false, 'boolean': 'foo', 'string': -42 }"); @@ -829,9 +830,21 @@ static void test_visitor_in_errors(TestInputVisitorData *data, g_assert(p->string == NULL); error_free(err); + err = NULL; g_free(p->string); g_free(p); visitor_input_teardown(data, NULL); + + v = visitor_input_test_init(data, "[ '1', '2', false, '3' ]"); + /* FIXME - a failed parse should not leave a partially-allocated + * array for us to clean up; this could cause callers to leak + * memory. */ + visit_type_strList(v, &q, NULL, &err); + assert(q); + assert(err); + qapi_free_strList(q); + error_free(err); + visitor_input_teardown(data, NULL); } int main(int argc, char **argv) -- 2.4.3