On Tue, Jun 07, 2016 at 04:42:38PM +0300, Jonathan Afek wrote:
> The BSS supplicant command returns a 0 len reply on a failure
> when the BSS isn't in the DB or if the command has invalid params.
It sounds fine to return FAIL on invalid params, but BSS-not-found is
not really a failure and I think the empty response for used here by
design. Changing that now could result in breaking existing users.
> This patch make the function wpa_supplicant_ctrl_iface_bss()
> return -1 on failures instead of 0 and therefore the BSS command
> will return FAIL on failures.
> This is important since when using udp control interfaces
> sending a 0 len reply does not send a reply at all and the
> control interface can wait forever (or until a timeout is reached)
> for a reply.
Maybe the UDP control interface backend should be fixed instead?
> diff --git a/wpa_supplicant/ctrl_iface.c b/wpa_supplicant/ctrl_iface.c
> @@ -4448,7 +4448,7 @@ static int wpa_supplicant_ctrl_iface_bss(struct wpa_supplicant *wpa_s,
> if ((ctmp = os_strchr(cmd + 6, '-')) == NULL) {
> wpa_printf(MSG_INFO, "Wrong BSS range "
> "format");
> - return 0;
> + return -1;
This looks fine..
> @@ -4530,7 +4530,7 @@ static int wpa_supplicant_ctrl_iface_bss(struct wpa_supplicant *wpa_s,
> }
>
> if (bss == NULL)
> - return 0;
> + return -1;
This could potentially break existing users that might expect an empty
response at the end of the iteration.
--
Jouni Malinen PGP id EFC895FA
_______________________________________________
Hostap mailing list
Hostap@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/hostap
