Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!mnetor!uunet!husc6!rutgers!ames!lll-tis!ptsfa!ihnp4!cbosgd!mandrill!hal!ncoast!allbery From: allbery@ncoast.UUCP (Brandon S. Allbery) Newsgroups: comp.os.minix Subject: HD problem finally fixed! and HD boot question Message-ID: <4240@ncoast.UUCP> Date: Sun, 23-Aug-87 00:48:52 EDT Article-I.D.: ncoast.4240 Posted: Sun Aug 23 00:48:52 1987 Date-Received: Wed, 26-Aug-87 06:42:16 EDT Reply-To: allbery@ncoast.UUCP (Brandon S. Allbery) Organization: Cleveland Public Access UN*X, Cleveland, Oh Lines: 87 I have finally gotten a chance to install the 1.2 patches. Actually, at present it's only the new klib88.s and xt_wini.c installed as yet, but the rest is ready to install. Some comments: (1) I finally convinced Minix to access the rest of my ST225. I don's know if this is an ITT XTRA BIOS bug or what, but when I gave up before I mentioned that Minix was acting as if I only had a 10MB hard drive. Guess what! That was EXACTLY the problem -- for some reason, it thinks I have a 10MB primary hard drive and a 20MB secondary. I swapped around the parameter tables after loading, and also altered the write precomp value (it said 128, my stuff says 256), both in wini.c (as opposed to xt_wini.c) and now have a 10MB partition for Minix to supplement the 5MB one. (2) I just looked at cal.c from the net a few months ago: at first glance, it didn't work because the stdio buffers weren't being flushed. I'll test it tomorrow and report back... (3) Speaking of which: why are the following not in the 1.2 commands package? cal cdiff ed expr test All of them were posted on the net. (4) What, officially, is the status of "patch"? I have successfully gotten it to run under Minix, via the patches posted to the net some time ago. "Fix" is OK unless you make changes to stuff on your own (such as my wini.c fixes...) (5) I still have plans to redo the tty stuff. It occurs to me that the changes to FS as proposed in an earlier fix may be a good idea; the current setup is stylistically nice but worthless as a useful construct. It teaches the use of message passing, but not how to write an OS that is actually useful. I *don't* have plans to work on that problem, though. (6) And, the biggie: Recently, Andy Tanenbaum posted a patch to Minix to load the root filesystem from the hard disk. This is nice, but I'd like to know if there will ever be an official way to have the root filesystem be ON the hard disk: a larger filesystem and more free memory are nice incentives. (Note that the ITT XTRA is an 8088 machine, not an 80286.) One thught I've had is to write a more conventional bootstrap loader, i.e.: Minix 1.2 boot : =hd(0,2)minix Loading hd(0,2)minix ... The '=' flags a boot command and allows the Olivetti M24 test; the rest tells the loader where to look for the program. Supported commands would be something like: =spec spec is: fd(n,n)file hd(n,n)file as "minix" or "msdos" ("ibmbio"?) or "ccpm" or etc., default extensions of "com" or "sys" (or none) -- where (n,n) is device and partition; for floppies the partition must be 0 dos boot the first bootable DOS partition minix boot the first unknown-type partition as Minix cpm boot from a partition of type 240 (CP/M-86, CCP/M, Concurrent DOS in native mode) m do an INT 18 (monitor under ITT XTRA, BASIC elsewhere) f fsck p fdisk This would be stored in a special partition consisting of a bootblock, a table of programs, and the programs themselves; the bootblock loads the first program, which is the actual loader. Any thoughts? -- Brandon S. Allbery, moderator of comp.sources.misc and comp.binaries.ibm.pc {{harvard,mit-eddie}!necntc,well!hoptoad,sun!mandrill!hal}!ncoast!allbery ARPA: necntc!ncoast!allbery@harvard.harvard.edu Fido: 157/502 MCI: BALLBERY <> ** Site "cwruecmp" is changing its name to "mandrill". Please re-address ** *** all mail to ncoast to pass through "mandrill" instead of "cwruecuecuom le