Dan Williams wrote:
> On Sun, 2008-11-16 at 10:01 +0000, Dave wrote:
>> 2. Does the driver need to send a dissociation event (or something) to
>> userspace on ifconfig down?
>
> If the association with the AP is no longer valid, then yes. Most of
> the other drivers do this already, I think.
I couldn't find in in mac80211, but a few (but not all) of the fullmac
drivers seem to do it.
Andrey: the attached patch may help, but given what you've written in
your other email (which I'll respond to separately), I don't think it
will fix things. Note that it's against wireless testing (+ private
changes). Should apply cleanly to an older tree after you account for
the file move.
Regards,
Dave.
---
diff --git a/drivers/net/wireless/orinoco/orinoco.c
b/drivers/net/wireless/orinoco/orinoco.c
index f8db026..007d5b8 100644
--- a/drivers/net/wireless/orinoco/orinoco.c
+++ b/drivers/net/wireless/orinoco/orinoco.c
@@ -2319,7 +2319,21 @@ int __orinoco_down(struct net_device *dev)
hermes_set_irqmask(hw, 0);
hermes_write_regn(hw, EVACK, 0xffff);
}
-
+
+ /* If associated, tell userspace we're dropping it */
+ if ((priv->last_linkstatus == HERMES_LINKSTATUS_CONNECTED) ||
+ (priv->last_linkstatus == HERMES_LINKSTATUS_AP_CHANGE) ||
+ (priv->last_linkstatus == HERMES_LINKSTATUS_AP_IN_RANGE))
+ {
+ union iwreq_data wrqu;
+
+ wrqu.data.length = 0;
+ wrqu.data.flags = 0;
+ wrqu.ap_addr.sa_family = ARPHRD_ETHER;
+ memset(wrqu.ap_addr.sa_data, 0, ETH_ALEN);
+ wireless_send_event(dev, SIOCGIWAP, &wrqu, NULL);
+ }
+
/* firmware will have to reassociate */
netif_carrier_off(dev);
priv->last_linkstatus = 0xffff;
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
