Date: Mon, 3 Jul 2000 18:01:09 +0100 (BST) From: Nick Hibma <n_hibma@calcaphon.com> To: Garrett Wollman <wollman@khavrinen.lcs.mit.edu> Cc: Warner Losh <imp@village.org>, Poul-Henning Kamp <phk@critter.freebsd.dk>, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/sys bus.h bus_private.h src/sys/kern subr_bus.c Message-ID: <Pine.BSF.4.20.0007031751380.21424-100000@localhost> In-Reply-To: <200007031634.MAA17071@khavrinen.lcs.mit.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
> In other words, devices which are removed need to be accounted for as > ``still attached but not physically present'' rather than simply > ``gone''. Could you give an example of when this would be good behaviour, because I can only see big problems when trying to implement this. In general when you willfully remove a device you want it to act like it has disappeared. For example the route over a USB ethernet dongle should be gone so you don't have to wait for timeouts on transfers, and get network down instead. In the case of an accidental remove of a zip drive, it might make sense to have the device appear at the same spot where you removed it. E.g. if the drive was mounted at that point in time. It however opens a whole can of worms: How do I reliably detect that a device is the same that went away before? What about the medium in it? The Linux USB boys discussed this at least two times and decided that it was not feasible. Allthough the USB devices can contain a serial number, it is not mandatory unfortunately. What if the user plugs the device into a different slot or port? Is it still the same? Nick -- n_hibma@webweaving.org n_hibma@freebsd.org USB project http://www.etla.net/~n_hibma/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.20.0007031751380.21424-100000>