Date: Sun, 24 Nov 1996 20:29:03 -0800 (PST) From: "Rodney W. Grimes" <rgrimes@GndRsh.aac.dev.com> To: darrylo@sr.hp.com Cc: bde@zeta.org.au, current@freefall.freebsd.org, jkh@time.cdrom.com Subject: Re: 2.2-ALPHA install failure Message-ID: <199611250429.UAA01410@GndRsh.aac.dev.com> In-Reply-To: <199611250142.AA145216177@hpnmhjw.sr.hp.com> from Darryl Okahata at "Nov 24, 96 05:42:57 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
> > > >> The data for partition 3 is: > > > >> sysid 165,(FreeBSD/NetBSD/386BSD) > > > >> start 0, size 50000 (24 Meg), flag 80 > > > >> beg: cyl 0/ sector 1/ head 0; > > > >> end: cyl 1023/ sector 63/ head 255 > > > > ^^ bogus (ie this means 62 sectors!!) > > Are you sure? All of the documentation and source code (e.g., > diskslice_machdep.c) that I've seen leads me to believe that the > "ending" sector is the address of the last sector in the partition. If > so, then the above means "63" sectors, not 62 (BIOS sector numbering > starts at 1, not 0). A fence post error on my part, (63 - 1) + 1 == 63, I forgot to count the fence post at ``1'' :-(. > > > >I would like to also see that corrected/changed to: > > > > The data for partition 3 is: > > > > sysid 165,(FreeBSD/NetBSD/386BSD) > > > > start 0, size 2097152 (1024 Meg), flag 80 > > > > beg: cyl 0/ sector 1/ head 0; > > > > end: cyl 1023/ sector 32/ head 63 > > > > > > This would not work so well. The historical_bogus_partition_table works > > > perfectly with ncr and other geometry-detecting controllers _because_ > > > it is perfectly invalid. Your table is just as invalid for disks smaller > > > than 1GB and for non-SCSI disk [drivers] that can't handle either 255/63 > > > or 64/32 geometries. > > > > At least my table will save your ass if this is sd1 and <1G and you decide > > to install Win95 or WinNT on your box, which will gladly partition the > > 1G-50MB in your case, AND FORMAT IT FOR YOU!!! > > Hey, "enhance your calm". ;-) Until now, no one's given a good > reason why it should be changed. I'd say that this is a pretty good > reason. I gave the same reason over a year ago, and at that time sysinstall was fixed to atleast cover this case, and now disklabel has been fixed to stop smashing the MBR table when installing boot1 to sector 0 so things have gotten far better. There are still cases that the bogus partition table ends up installed, and we need to eliminate those. > > However, if it's changed, we'll then have two > "historical_bogus_partition_tables" (the old one and the new one). Or a ``very dangerous one'' and a ``not so dangerous one''... :-) > > > >Or atleast make size == cyl * sec * head (ie 1024*64*255 == 16711680) > > > >my reasoning is that any third party fdisk program that sees the current > > > >bogus table likes to try and ``fix it'' and usually ends up doing quite > > > >the wrong thing : -(. > > Well, if the partition table must be bogus for geometry-detecting > controllers, how about changing the size to 2^32-1? Do the > geometry-detecting controllers properly detect this? I'd try this > myself, but I don't have the time. I'll give that a test here in a minute, about to trash a 4.5G drive for some testing any way.... (I only be testing on an NCR 53C810, so it won't give us much data....) -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation, Inc. Reliable computers for FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199611250429.UAA01410>