Date: Mon, 24 Jan 2000 12:44:53 +0800 From: Greg Lehey <grog@lemis.com> To: Assar Westerlund <assar@sics.se>, Brian Beattie <beattie@aracnet.com> Cc: Robert Watson <robert+freebsd@cyrus.watson.org>, fs@FreeBSD.ORG Subject: Re: UDF, userfs Message-ID: <20000124124453.G2643@mojave.worldwide.lemis.com> In-Reply-To: <5liu0l1ay3.fsf@assaris.sics.se>; from assar@sics.se on Sun, Jan 23, 2000 at 08:52:52AM %2B0100 References: <Pine.LNX.4.10.10001230117530.4117-100000@shell1.aracnet.com> <5l1z79um0a.fsf@assaris.sics.se> <5liu0l1ay3.fsf@assaris.sics.se> <Pine.LNX.4.10.10001230117530.4117-100000@shell1.aracnet.com> <Pine.LNX.4.10.10001211012020.28236-100000@shell1.aracnet.com> <Pine.BSF.3.96.1000121184917.66083B-100000@fledge.watson.org> <20000122131656.C391@mojave.worldwide.lemis.com> <5liu0l1ay3.fsf@assaris.sics.se>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sunday, 23 January 2000 at 8:52:52 +0100, Assar Westerlund wrote: > Greg Lehey <grog@lemis.com> writes: >> Hmm. A kld runs in kernel context, not user context. Sure, it's >> easier to load than rebuilding a kernel, and I believe klds are the >> correct approach to added kernel functionality, but it doesn't offer >> one of the prime advantages of userland development: if your program >> crashes, your program crashes, not the system. If you're developing a >> kld, a bug can crash the system. > > Yes, but both the Coda and the Arla kld are very simple and all the > real work (and thus, the devlopment) takes part in the user space > daemon. The kld is mostly there as a way of communicating with the > kernel. They work now. How long did it take to debug them? On Sunday, 23 January 2000 at 11:21:09 +0100, Assar Westerlund wrote: > Brian Beattie <beattie@aracnet.com> writes: >>> Yes, but both the Coda and the Arla kld are very simple and all the >> >> I would disagree that they are very simple. > > Well, make that `simple and much simpler than the corresponding code > inside the kernel would be' then? The point is that you're keeping > most of the complexity and the code that changes outside of the > kernel. It sounds like what you're saying is that the kld is an interface to userland code which does work which would normally be in the kernel. That's different from most klds, which contain code identical to that in the kernel. That might be an option, but in fact the old block device interface was exactly such an option. My suggestion to have it as a kld appears to be pretty much the same concept as what you're talking about, though I suspect the interface would be at a lower level. Greg -- Finger grog@lemis.com for PGP public key See complete headers for address and phone numbers To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20000124124453.G2643>