Return-Path: Received: by vger.rutgers.edu id <970791-11449>; Mon, 9 Mar 1998 02:36:38 -0500 Received: from warthog.nexor.co.uk ([128.243.9.237]:2132 "EHLO nexor.co.uk" ident: "SOCKWRITE-65") by vger.rutgers.edu with ESMTP id <970814-11449>; Mon, 9 Mar 1998 02:36:11 -0500 To: "William R. Kerr" Cc: linux-kernel@vger.rutgers.edu Subject: Re: cPCI Hot Swap for Linux (system-level issues) In-reply-to: Your message of "Wed, 04 Mar 1998 22:53:47 PST." <199803050653.WAA32072@ridge.spiritone.com> Date: Mon, 09 Mar 1998 08:57:05 +0000 From: David Howells Message-Id: <19980309073624Z970814-11449+4748@vger.rutgers.edu> X-Envid: [/PRMD=NEXOR/ADMD= /C=GB/;nexor.co.uk:289350:980309085725] Sender: owner-linux-kernel@vger.rutgers.edu Content-Length: 1164 Lines: 26 > All issues of this type can be resolved, and I would urge people with more > recent experience in kernel workings than I to begin drafting some plans for > how this could be cleanly laid out. The goal would be write a portable "bus > driver" and a set of clearly specified conventions that writers of device > drivers could follow so that drivers could be written that would work both > with normal PCI implementations of the device as well as with Hot Swap cPCI > implementations. My configuration manager was written with this sort of thing in mind (though I was thinking in terms of PnP-BIOS hot-swap at the time)... It uses automatic rebinding of resources (with driver notification) as its method of conflict resolution. However, I haven't gone through and modified the majority of the vast collection of drivers that Linux has available, as (1) I can't test most of them, and (2) I would have to re-modify them every time I ported to a new kernel version. See: http://lucifer.hemmet.s-hem.chalmers.se/~dwh David Howells - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu