Hi Joey, On Tue, Oct 19, 2010 at 01:45:00AM -0600, Joey Lee wrote: > Hi Samuel, > > 於 五，2010-10-15 於 00:48 +0200，Samuel Ortiz 提到： > > > I also attached it on > > > http://bugs.meego.com/show_bug.cgi?id=8075 > > So I've been looking at this issue and this race condition is indeed real, > > and > > not related to your specific hardware. It will be visible to HW that is > > either > > removable (e.g. a US dongle) or completely removed from the bus upon rfkill > > commands (e.g. pcie hotpluggable devices). > > When the device shows up through udev/rtnl before wpa_supplicant is started, > > the corresponding element->driver pointer will not be set, and we'll be left > > with a stale connman_device upon removal of the physical device.. > > > > Although your patch is correct and does fix the issue, I don't really like > > exporting the whole element probing routine, that should be kept internal to > > the element code. > > > > What I propose is to create a simple __connman_element_set_driver() routine > > that will just set the right driver for a given element. The patch below > > works > > fine for me, could you please give it a try (It may not apply cleanly > > depending on the ConnMan version you're using) ? > > > > Thank's for your review and patch. > > This patch works to me on MSI netbook can fix the wifi hotkey issue, it > used pcie hot plug for wifi module. Thanks for testing it. Patch is now applied.
Cheers, Samuel. -- Intel Open Source Technology Centre http://oss.intel.com/ _______________________________________________ connman mailing list [email protected] http://lists.connman.net/listinfo/connman
