Hi Daniel,
May i know if you a chance to look the logs below please?

Deepu Paul

On Wed, Jan 8, 2020 at 5:28 PM Deepu Paul <pauldeepu16@gmail.com> wrote:
Hi Daniel,

Attaching both connman and wpa_supplicant logs,

There are 2  disconnects with reason code 1, in the first one it reconnects as expected, in the second one it didn't.

I am bit confused here, not sure if this an issue with wpa_supplicant.

Can you please help me with the logs to find the root cause?

Deepu Paul

On Tue, Jan 7, 2020 at 8:04 AM Daniel Wagner <wagi@monom.org> wrote:
Hi Deepu,

On Mon, Jan 06, 2020 at 12:22:31PM +0000, Deepu Paul wrote:
> Hi Daniel,
> The previous mail got bounced back as the attachment  size was greater than
> 500kb .
> Attaching the logs again.

Thanks for the logs. Really helpful.

> one thing i am noticing is every-time during disconnect, there is
> a (Reason: 3=DEAUTH_LEAVING)[line 6790] during the re association
> after (Reason: 1=UNSPECIFIED)[line 5752].

IIRC, deauth leaving means the local device has decided to leave the
network. Though this doesn't make any difference for ConnMan. It's the
same code path. Only reason code 6 is handled differently.

> I am not sure if this is the issue with wpa_supplicant, can you please
> check the logs and let me know if this is initiated by the wpa_supplicant?

From you previous email:

> In the second one it failed, in this scenario i don't see the interface state change to 2(disconnected), in this case the state  change is (4, 5, 6, 8, 10)
> and seeing the message  "Probably roaming right now! Staying connected..."

That is what I suspected. So wpa_supplicant doesn't tell us excactly
what it is doing. Suddenldy we see state transitions and we have to
guess what's happending.

But then we get a new message which it the disconnect reason:

connmand[3595]: ../connman-1.37/src/rtnl.c:rtnl_message() NEWLINK len 72 type 16 flags 0x0000 seq 0 pid 0
connmand[3595]: ../connman-1.37/gsupplicant/supplicant.c:signal_interface_changed()
connmand[3595]: ../connman-1.37/gsupplicant/supplicant.c:interface_property() DisconnectReason
connmand[3595]: ../connman-1.37/gsupplicant/supplicant.c:signal_bss_removed()
connmand[3595]: ../connman-1.37/gsupplicant/supplicant.c:update_network_signal() New network signal -90
connmand[3595]: ../connman-1.37/plugins/wifi.c:network_removed() name purpleline

Could it be, whenever it works you don't get an disconnect reason
after we assume wpa_supplicant is roaming?

connman mailing list -- connman@lists.01.org
To unsubscribe send an email to connman-leave@lists.01.org