On Mon, 19 Apr 1999 takaw...@shidahara1.planet.sci.kobe-u.ac.jp wrote: > In message <pine.bsf.3.96.990418205350.3261e-100...@heidi.plazza.it>, Nick > Hibm > a wrote: > >> Why would I say it wasn't ready? Because outside of core (apparently), > >> nobody was warned/told that this was going to be committed in a few > >> days/hours/minutes. > > >I've ported the newconfig style USB code of NetBSD to FreeBSD and I > >really prefer the newbus style, because it is truly dynamic. > > Simple Question. > If there were 'Closed'-Host-Controller-Interface with object-only driver, > Can the vendor make the Host controller recognized without changing > usb.c code? > > #That's what frustrated me while writing driver for smbus controller.
It really depends on the nature of the bus. For plug-and-play style busses such as USB and PCI, the controller can decide what devices exist without the help of drivers and can then uniquely match them with the right driver. This trivially allows drivers to be added after the system is running. For 'manual' style busses, the creation of devices is generally controlled by hints. To create a new device for a new driver, it would be necessary to add the hint information for the new device and prompt the controller to add it. This isn't quite there yet but will be needed soon for dynamically adding devices (e.g. /usr/bin/joy). -- Doug Rabson Mail: d...@nlsystems.com Nonlinear Systems Ltd. Phone: +44 181 442 9037 To Unsubscribe: send mail to majord...@freebsd.org with "unsubscribe freebsd-current" in the body of the message
