On Wed, Feb 12, 2020 at 22:10:42 +0800, Yi Wang wrote:

From: Huang Zijiang <huang.zijiang@zte.com.cn>

virJSONValueObjectGetObject maybe return NULL if the key is missing or if value is not the correct TYPE, so we have to prevent a NULl pointer from being accessed.

Signed-off-by: Huang Zijiang <huang.zijiang@zte.com.cn> Signed-off-by: Yi Wang <wang.yi59@zte.com.cn> --- src/util/virqemu.c | 5 +++++ 1 file changed, 5 insertions(+)

diff --git a/src/util/virqemu.c b/src/util/virqemu.c index f3a233a..29fbe4e 100644 --- a/src/util/virqemu.c +++ b/src/util/virqemu.c @@ -257,6 +257,11 @@ virQEMUBuildObjectCommandlineFromJSON(virBufferPtr buf, const char *type = virJSONValueObjectGetString(objprops, "qom-type"); const char *alias = virJSONValueObjectGetString(objprops, "id"); virJSONValuePtr props = virJSONValueObjectGetObject(objprops, "props"); + if (!props) { + virReportError(VIR_ERR_INVALID_ARG, "%s", + _("reply was missing return props data"));

This error message is misleading. There's no reply, but rather this is an internally generated JSON struct

+ return -1; + }

return virQEMUBuildObjectCommandlineFromJSONInternal(buf, type, alias, props);

This function checks 'props' internally. What is this patch supposed to fix?