Date: Tue, 15 Oct 2002 06:36:52 +1000 From: John Birrell <jb@cimlogic.com.au> To: Josef Karthauser <joe@FreeBSD.org> Cc: John Birrell <jb@cimlogic.com.au>, arch@FreeBSD.org Subject: Re: USB driver design issues Message-ID: <20021015063651.I261@freebsd1.cimlogic.com.au> In-Reply-To: <20021014170949.GB25180@genius.tao.org.uk>; from joe@FreeBSD.org on Mon, Oct 14, 2002 at 06:09:49PM %2B0100 References: <20021014083622.G261@freebsd1.cimlogic.com.au> <20021014170949.GB25180@genius.tao.org.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Oct 14, 2002 at 06:09:49PM +0100, Josef Karthauser wrote: > There are known bugs in -stable that I imagine won't ever get fixed. > The stack has been reengineered in NetBSD though and ported to > -current, but it remains to be seen whether anyone's got the time > to port it to -stable. I might try that. A potential problem is that I only have a few USB devices. > I don't know about the issue that you mention particularly, but it does > appear that device disconnection is now very stable in -current. It may > be worth trying to put a current box together to try it, or finding > someone who can test your driver on a -current box. No it doesn't (on inspection). The uhub driver is missing DEVMETHOD(bus_child_detached, uhub_child_detached) in the uhubroot so that device isn't cleared when a driver is kldunloaded. The -current version has DEVMETHOD(bus_driver_added, uhub_driver_added), but the method is empty. This is required in -stable to stop the bus_generic_driver_added from being called and probing the devices without setting up the attach_arg structure. And empty uhub_driver_added method means that a USB device has to be disconnected and re-connected before the device is probed as ugen after a driver is kldunloaded. -- John Birrell To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-arch" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20021015063651.I261>