Den 1 aug 2014 13:01 skrev "Patrik Flykt" <patrik.fl...@linux.intel.com>: > > In normal cases when an error reply is received from wpa_supplicant, > the code uses a DBusMessageIter struct from the stack and initializes > it to point to the error return message in gsupplicant/dbus.c, > method_call_reply(). When passed to parse_supplicant_error(), the > iterator is valid, but contains no data and everything works fine. > > When a gsupplicant pending call is cancelled by ConnMan, the > cancellation code will call the callback instead with a NULL iterator. > Explicitely catch this NULL iterator and return the default -ECANCELED > to the caller instead of relying on the specific way the dbus library > was compiled - either detecting NULL pointers or just plainly crashing. > > The fix is based on a very similar one by Rickhard Röjfors but was made > to be even more explicit and accompanied by a longer explanation.
Except from that my name was incorrectly spelled, I ack this. A stack trace says more than 1000 words ;-) > --- > gsupplicant/supplicant.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/gsupplicant/supplicant.c b/gsupplicant/supplicant.c > index d26b6e2..1886a40 100644 > --- a/gsupplicant/supplicant.c > +++ b/gsupplicant/supplicant.c > @@ -3792,6 +3792,9 @@ static int parse_supplicant_error(DBusMessageIter *iter) > int err = -ECANCELED; > char *key; > > + if (!iter) > + return err; > + > /* If the given passphrase is malformed wpa_s returns > * "invalid message format" but this error should be interpreted as > * invalid-key. > -- > 1.9.1 > > _______________________________________________ > connman mailing list > connman@connman.net > https://lists.connman.net/mailman/listinfo/connman _______________________________________________ connman mailing list connman@connman.net https://lists.connman.net/mailman/listinfo/connman
