Regression with ndis and rt2860 on 8.0-RC1

[Paul B. Mahol]

On 9/24/09, Paul B. Mahol <onemda at gmail.com> wrote:
> On 9/24/09, Christian Laursen <xi at borderworlds.dk> wrote:
>> Christian Laursen wrote:
>>> Paul B. Mahol wrote:
>>>> On 9/23/09, Paul B. Mahol <onemda at gmail.com> wrote:
>>>>> On 9/22/09, Christian Laursen <xi at borderworlds.dk> wrote:
>>>>>> Hi,
>>>>>>
>>>>>> Some days ago I upgraded my Asus Eee 1000H from 7.2-RELEASE to
>>>>>> 8.0-BETA4
>>>>>> and then to 8.0-RC1 as of this morning.
>>>>>>
>>>>>> Since the Ralink Wireless NIC is not supported with a native driver I
>>>>>> have been using ndis for a long time on 7.x without any issues.
>>
>> I just made a discovery.
>>
>> If I create a wlan device with wlandevice ndis0, I am able to associate
>> and get traffic through the interface.
>>
>> I did get one warning though:
>> wlan0: ieee80211_new_state_locked: pending SCAN -> RUN transition lost
>>
>> If I read UPDATING correctly this is the way to configure wireless
>> devices on 8.0. As this is my first 8.0 system I was not sufficiently
>> aware of that.
>>
>> However, a panic is somewhat unexpected in this case.
>
> Yes, there is still code path where vap is used when it is NULL.

What about moving windrv_(un)wrap() calls from ndisdrv_modevent() to
vap create/delete.

I dont like:
"if (vap == NULL) then return" solution.

-- 
Paul