**** BEGIN LOGGING AT Wed May 05 02:59:56 2010
May 05 08:27:54 <holtmann> zhenhua: pong
May 05 08:45:25 <zhenhua> holtmann: i am investigating a bug #395. it's related with udev. but i think it's more like a udev bug instead of ofono issue.
May 05 08:45:28 <zhenhua> holtmann: http://bugs.meego.com/show_bug.cgi?id=395
May 05 08:46:29 <zhenhua> holtmann: it only occurs on netbook. i cannot reproduce it on ubuntu 9.10 or 10.4
May 05 08:46:54 <holtmann> What kernel version?
May 05 08:47:26 <holtmann> the USB ACM driver had a bug that caused the ttyACMx not to go away. Reference counting issue.
May 05 08:48:13 <zhenhua> kernel 2.6.33.2-8
May 05 08:49:04 <zhenhua> but huawei modem also could not go away. the parent node reported from udev was wrong.
May 05 08:49:34 <zhenhua> what we expected is /devices/pci0000:00/0000:00:1d.7/usb2/2-1
May 05 08:49:40 <zhenhua> but returned parent is /devices/pci0000:00/0000:00:1d.7/usb2
May 05 08:50:35 <zhenhua> udev does report device remove event to us
May 05 09:39:48 <zhenhua> holtmann: one workaround to fix #395 is to change ofono logic so that we don't not depends on parent node.
May 05 09:40:17 <zhenhua> holtmann: instead, we record the parent node as user data of struct *udev
May 05 09:40:31 <zhenhua> holtmann: what do u think?
May 05 09:49:29 <holtmann> I am not following. You mean inside the ConnMan plugin, right?
May 05 09:49:38 <holtmann> Or inside oFono?
May 05 09:50:14 <zhenhua> in ofono
May 05 09:50:33 <zhenhua> plugins/udev.c
May 05 09:51:36 <zhenhua> currently it searchs three of parent nodes and try to find OFONO_DRIVER property.
May 05 10:19:29 <holtmann> And what is the problem with that?
May 05 10:21:11 <zhenhua> holtmann: it cannot find OFONO_DRIVER property now, because the first parent is already wrong.
May 05 10:22:03 <zhenhua> holtmann: go home now. talk to u later.
May 05 10:22:24 <holtmann> I see. So we need something a bit smarter. Do you have a patch that fixes this (while not breaking MBM and HSO).
May 05 10:23:20 <zhenhua> holtmann: i am trying to make one now and i will ask u to review tomorrow.
May 05 10:24:55 <holtmann> So problem area here is that in some cases we have to go to a parent to find the driver.
May 05 10:25:21 <holtmann> And in some other cases where a netdev and a TTY is needed, we have to have the parent of a parent information and use them.
May 05 23:20:40 <nt> hi
**** ENDING LOGGING AT Thu May 06 02:59:56 2010