From owner-cvs-all Fri Aug 20 3:59: 0 1999 Delivered-To: cvs-all@freebsd.org Received: from overcee.netplex.com.au (overcee.netplex.com.au [202.12.86.7]) by hub.freebsd.org (Postfix) with ESMTP id E5EA214DF4; Fri, 20 Aug 1999 03:58:50 -0700 (PDT) (envelope-from peter@netplex.com.au) Received: from netplex.com.au (localhost [127.0.0.1]) by overcee.netplex.com.au (Postfix) with ESMTP id 8D3271C1F; Fri, 20 Aug 1999 18:57:38 +0800 (WST) (envelope-from peter@netplex.com.au) X-Mailer: exmh version 2.0.2 2/24/98 To: Poul-Henning Kamp Cc: Peter Jeremy , julian@whistle.com, cvs-all@FreeBSD.org, cvs-committers@FreeBSD.org, grog@lemis.com Subject: Re: Splitting struct buf In-reply-to: Your message of "Fri, 20 Aug 1999 07:01:45 +0200." <2046.935125305@critter.freebsd.dk> Date: Fri, 20 Aug 1999 18:57:38 +0800 From: Peter Wemm Message-Id: <19990820105738.8D3271C1F@overcee.netplex.com.au> Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk Poul-Henning Kamp wrote: > In message <99Aug20.101641est.40337@border.alcanet.com.au>, Peter Jeremy writ es: > >Julian Elischer wrote: > >>changing both is also a fair idea. > >>(that way we know that someone looked at ALL the places the present struct > >>buf is used.. :-) > > > >Ignoring the smiley, I think this is probably the best suggestion. It > >provides a clear `heads-up' for any independent device writers that > >the usage has changed. There are about 750 references to struct buf > >in the kernel - missing one would be quite easy. > > > >If only one name changes, then POLA would suggest that `struct buf' > >remain associated with I/O requests (which is the historical and > >probably most common usage). > > I'm happy either way, I'm not religiously attached to any of the > names. For what it's worth, given all the options, I'd prefer that we gutted all the caching aspects from 'struct buf' and rename the caching part to something else. The reason being, the impact of the caching part is confined to source that we control and maintain, while the IO part is very heavily exposed to the device drivers. I don't think we gain much by turning the device driver interface totally on it's head and making it harder still to port drivers (not only relative to other unixen, but older freebsd versions too). Of course, there are other options too, but if we can avoid making driver portability 10 times harder without making it too inconvenient for us, I think it's worth doing. Cheers, -Peter -- Peter Wemm - peter@FreeBSD.org; peter@yahoo-inc.com; peter@netplex.com.au To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message