qdev: Reject drive property override

qdev_prop_set_drive() screws up when the property already has a
non-null value: it neglects to release the old value.  Both the old
and the new backend become attached to the same device.

Example (taken from iotest 172): -fda ... -drive if=none,... -global
floppy.drive=none0.

Special case: attempting to use the same backend both times fails.
Example (also from iotest 172): -fda ... -global floppy.drive=floppy0.

Yet another example: -device with multiple drive=... (but not
device_add, which silently drops all but the last duplicate property).

Perhaps drive property override could be made to work.  Perhaps it
should.  I can't afford the time to figure this out now.  What I can
do is reject usage that leaves backends in unhealthy states.  For what
it's worth, we've long done the same for netdev properties.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20200622094227.1271650-12-armbru@redhat.com>

1 files changed, 8 insertions, 0 deletions

diff --git a/hw/core/qdev-properties-system.c b/hw/core/qdev-properties-system.c
index 3f84309b7e..6b5fc59901 100644
--- a/hw/core/qdev-properties-system.c
+++ b/hw/core/qdev-properties-system.c
@@ -98,6 +98,14 @@ static void set_drive_helper(Object *obj, Visitor *v, const char *name,
         return;
     }
 
+    /*
+     * TODO Should this really be an error?  If no, the old value
+     * needs to be released before we store the new one.
+     */
+    if (!check_prop_still_unset(dev, name, *ptr, str, errp)) {
+        return;
+    }
+
     if (!*str) {
         g_free(str);
         *ptr = NULL;