ap_vlan_open exits hwsim tests

Jacob Minshall jacob at cozybit.com
Wed Mar 12 03:43:33 EDT 2014


I have found the offending commit


89c86a64cd056e283323710c9ddf6f7090a450c8
kobject: delay kobject release for random time

I guess it is doing what is says, and is indeed taking a random amount
of time (which just happens to be more time than the vm is willing to
wait)

So is extending the time of the wait to 20 a good solution? I'm not
personally fond of arbitrary wait times being extended as needed.
Though this patch does not seem wrong in and of itself.

On Tue, Mar 11, 2014 at 6:04 PM, Jacob Minshall <jacob at cozybit.com> wrote:
> On Tue, Mar 11, 2014 at 5:43 PM, Eduardo Abinader
> <eduardo.abinader at openbossa.org> wrote:
>> It was a timeout issue in wpaspy:
>>
>>     def request(self, cmd):
>>            self.s.send(cmd)
>>            [r, w, e] = select.select([self.s], [], [], 20)
>>            if r:
>>
>> I just increased the timeout and everything went fine. Don't know why do I
>> need now to do such change, in order to get my tests running ok. A change in
>> hwsim? I was using latest wireless-testing head for my tests.
>>
>> Jouni, it is ok if I send you a patch with this incremented value (from 10
>> to 20)? I've tested with 15 and it failed also.
>>
>> Cheers.
>>
>>
>> On Tue, Mar 11, 2014 at 3:53 PM, Eduardo Abinader
>> <eduardo.abinader at openbossa.org> wrote:
>>>
>>> I've run tests on kvm with v3.14-rc4 and on non-virtual machines. The
>>> behavior I reported, termination earlier, was due to a crash in my kvm. It
>>> occurs that this test is freezing my real machine, event if it is run
>>> separetely (run-tests.py -f ap_vlan).
>>>
>>> I suspect of configure filter function in hwsim that is causing such
>>> issue, as can be seen in the message logs below:
>>>
>>> ===> From VM logs:
>>> [ 1610.660319] kobject: 'brport' (ffff880078d35558): kobject_uevent_env
>>> [ 1610.660341] kobject: 'brport' (ffff880078d35558): kobject_uevent_env:
>>> filter function caused the event to drop!
>>> [ 1610.663776] kobject: 'rx-0' (ffff88007afc3b20): kobject_release, parent
>>> ffff88007afc1750 (delayed 100)
>>> [ 1610.667599] kobject: 'tx-3' (ffff880078df4bf8): kobject_release, parent
>>> ffff88007afc1750 (delayed 300)
>>> [ 1610.671548] kobject: 'tx-2' (ffff880078df49b8): kobject_release, parent
>>> ffff88007afc1750 (delayed 100)
>>> [ 1610.674346] kobject: 'tx-1' (ffff880078df4778): kobject_release, parent
>>> ffff88007afc1750 (delayed 400)
>>> [ 1610.675251] kobject: 'tx-0' (ffff880078df4538): kobject_release, parent
>>> ffff88007afc1750 (delayed 300)
>>>
>>> ===> From real machine logs:
>>> kernel: [  298.329711] ieee80211 phy2: mac80211_hwsim_config
>>> (freq=2412(2412 - 0)/noht idle=0 ps=0 smps=static)
>>> kernel: [  298.329728] ieee80211 phy2: mac80211_hwsim_configure_filter
>>> kernel: [  298.329744] IPv6: ADDRCONF(NETDEV_UP): wlan2: link is not ready
>>> kernel: [  298.329747] ieee80211 phy2: mac80211_hwsim_configure_filter
>>> kernel: imklog 5.8.11, log source = /proc/kmsg started.
>>>
>>>
>>> The kvm logs are available at
>>> https://www.dropbox.com/s/miuo7z2rscj2rwg/kvm_logs.tgz
>>>
>>>
>>> Thanks.
>>>
>>> On Tue, Mar 11, 2014 at 9:54 AM, Jouni Malinen <j at w1.fi> wrote:
>>>>
>>>> On Tue, Mar 11, 2014 at 09:13:57AM -0400, Eduardo Abinader wrote:
>>>> > With current master branch (390b92913a9a1b3a6aaf70e8b5971a7b7c76cabc),
>>>> > in
>>>> > every environment I
>>>> > test, hwsim test is exiting just after ap_vlan_open with the message:
>>>> > 'Terminating early due to device reset failure', although I have
>>>> > noticed
>>>> > success in buildbot.
>>>>
>>>> I don't see that in any of my test setups. Could you please check the
>>>> end of each ap_vlan_open log file to see if they show any signs of a
>>>> process getting stuck or terminating? If no, please check all log files
>>>> in the same directory that are newer than ap_vlan_open* logs.
>
> I have noticed the same behaviour when running in a virtual machine.
> The test will fail for me with the latest wireless-testing kernel, but
> it passes with v3.13 for me. I will do a git bisect tomorrow if no one
> else has by then. Since increasing the timeout fixes it, maybe
> whatever is happening is now taking longer than it did before?
>>>>
>>>> --
>>>> Jouni Malinen                                            PGP id EFC895FA
>>>> _______________________________________________
>>>> HostAP mailing list
>>>> HostAP at lists.shmoo.com
>>>> http://lists.shmoo.com/mailman/listinfo/hostap
>>>
>>>
>>
>>
>> _______________________________________________
>> HostAP mailing list
>> HostAP at lists.shmoo.com
>> http://lists.shmoo.com/mailman/listinfo/hostap
>>
>
>
>
> --
> Jacob Minshall
> cozybit Inc.
> http://www.cozybit.com



-- 
Jacob Minshall
cozybit Inc.
http://www.cozybit.com


More information about the HostAP mailing list