Date: Tue, 28 May 2002 09:58:55 -0700 From: Marcel Moolenaar <marcel@xcllnt.net> To: Poul-Henning Kamp <phk@critter.freebsd.dk> Cc: cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/conf NOTES files options src/sys/geom geom_gpt.c geom_mbr.c src/sys/sys gpt.h Message-ID: <20020528165855.GA306@dhcp01.pn.xcllnt.net> In-Reply-To: <597.1022589933@critter.freebsd.dk> References: <200205280904.g4S94nu58457@freefall.freebsd.org> <597.1022589933@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, May 28, 2002 at 02:45:33PM +0200, Poul-Henning Kamp wrote: > In message <200205280904.g4S94nu58457@freefall.freebsd.org>, Marcel Moolenaar w > rites: > > > The correct behaviour of the MBR class would be to back-off (reject) > > a MBR if it's a Protective MBR (a MBR with a single partition of type > > 0xEE that spans the whole disk (as far as the MBR is concerned). > > The correct behaviour if the GPT class would be to back-off (reject) > > a GPT if there's a MBR that's not a Protective MBR. > > I have no really good ideas for how to resolve the various and > conflicting methods claims to the same disk-image. I guess in the > end it will come down to a number of more or less crude heuristics > in the relevant methods. I think it would help to have a way to order the tasting. The MBR class tasted the disk first and claimed the partitions. The EFI spec says that you have to look for GPT first. In this case, the GPT class has to decide if it wants to treat the disk as a GPT disk or a MBR disk. I would then also help if a claim for a disk could stop further tasting. That way you don't have to do the whole song and dance of creating geoms, tasting and spoiling them again and have GEOM cleanup the duplicate geoms for the same partition (at least that's what I think is happening) > There is as such no harm in having both an MBR and a GPT on the > same device, as long as people do not use the wrong one. The question then becomes: what is the right one? The current EFI spec doesn't say much about having both, but it appears to be a problem in practice (using MBR tools to update the partitions and then be surprised that the GPT based OS doesn't see the changes). > It would be really beneficial if you can help me with producing a > XML disk image to put in the regression tests. Sure. I believe all the code is in the regression suite to make the xml files, but I've yet to play with it. I did look at it yesterday, but it was too late to really give it a serious effort. I'll do so today after work. I'll bug you if I get stuck :-) -- Marcel Moolenaar USPA: A-39004 marcel@xcllnt.net 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?20020528165855.GA306>