[comp.os.minix] The HD strikes again!

allbery@ncoast.UUCP (Brandon S. Allbery) (05/17/87)

Well, my replacement floppy drive finally arrived, and I've spent all after-
noon recompiling MINIX with the new xt_wini.c.  Summary:  Well, almost.

MINIX will now mkfs /dev/hd2 without errors.  (/dev/hd1 is my MS-DOS partition
and I have to leave it there.)  /dev/hd3 gets an error and quits; /dev/hd4
gets a whole slew of errors -- and sounds as if it's ramming the drive head
against the stop.  I configured all partitions as "DOS (1) Bootable" under
the EDPART.EXE program; changing this did not help.  I also can't assemble
the program to do this under MS-DOS, because while I have MASM, it's version
1.25 which is too buggy to do much of anything.

Partition information:

Drive: ST225, Controller: WD1002-WX2 without autoconfig
	17 sectors/track, 4 heads, 611 cylinders (512 bytes/sector)
Partition 1:	cylinders 000-150, 10267 sectors (DOS, bootable)
Partition 2:	cylinders 151-301, 10268 sectors, 5134 blocks
Partition 3:	cylinders 302-452, 10268 sectors, 5134 blocks
Partition 4:	cylinders 453-609, 10676 sectors, 5338 blocks
	(cylinder 610 is reserved for head parking)

I am at a loss for this.  Presumably the driver works, I know the drive is
working right (I reformatted before doing this); has anyone used EDPART.EXE
yet?  Is it perhaps broken?

I really only need 3 partitions (DOS, MINIX boot, /usr -- and the sources will
be in /usr/src).

Thanks in advance for any hints,
++Brando
-- 
Copyright (C) 1987 Brandon S. Allbery.  Redistribution permitted only if the
	redistributor permits further redistribution.
		 ---- Moderator for comp.sources.misc ----
Brandon S. Allbery	{decvax,cbatt,cbosgd}!cwruecmp!ncoast!allbery
Tridelta Industries	{ames,mit-eddie,talcott}!necntc!ncoast!allbery
7350 Corporate Blvd.	necntc!ncoast!allbery@harvard.HARVARD.EDU
Mentor, OH 44060	+01 216 255 1080	(also eddie.MIT.EDU)

allbery@ncoast.UUCP (Brandon Allbery) (05/18/87)

As quoted from <2513@ncoast.UUCP> by allbery@ncoast.UUCP (Brandon S. Allbery):
+---------------
| MINIX will now mkfs /dev/hd2 without errors.  (/dev/hd1 is my MS-DOS partition
| and I have to leave it there.)  /dev/hd3 gets an error and quits; /dev/hd4
| gets a whole slew of errors -- and sounds as if it's ramming the drive head
| against the stop.  I configured all partitions as "DOS (1) Bootable" under
+---------------

More info:  I cut it down to the MS-DOS partition (150 sectors) and a Minix
partition, using a different program (DOS 3.1 FDISK; the Minix partition was
created as a "second MS-DOS partition").  This will mkfs to 5304 blocks, but
anything higher fails.  MS-DOS successfully formats the entire disk, both
low-level and DOS FORMAT.  This is using the latest version of the xt_wini.c
patches (the last two differed only in the existence of autoconfig patches in
one of them, I do not have an autoconfig ROM).

I am totally bewildered, since everyone else seems to be having success with
the WD1002-WX2 and these patches.

++Brando
-- 
Copyright (C) 1987 Brandon S. Allbery.  Redistribution permitted only if the
	redistributor permits further redistribution.
		 ---- Moderator for comp.sources.misc ----
Brandon S. Allbery	{decvax,cbatt,cbosgd}!cwruecmp!ncoast!allbery
Tridelta Industries	{ames,mit-eddie,talcott}!necntc!ncoast!allbery
7350 Corporate Blvd.	necntc!ncoast!allbery@harvard.HARVARD.EDU
Mentor, OH 44060	+01 216 255 1080	(also eddie.MIT.EDU)

rmtodd@uokmax.UUCP (05/19/87)

In article <2513@ncoast.UUCP>, allbery@ncoast.UUCP (Brandon S. Allbery) writes:
> MINIX will now mkfs /dev/hd2 without errors.  (/dev/hd1 is my MS-DOS partition
> and I have to leave it there.)  /dev/hd3 gets an error and quits; /dev/hd4
> gets a whole slew of errors -- and sounds as if it's ramming the drive head
> against the stop.  I configured all partitions as "DOS (1) Bootable" under
> the EDPART.EXE program; changing this did not help.  I also can't assemble
> the program to do this under MS-DOS, because while I have MASM, it's version
> 1.25 which is too buggy to do much of anything.
If I remember aright, EDPART by default creates the partitions as DOS
extension (4), which works fine for MINIX.  As for the program supplied
on the MINIX disks, I haven't tried it--it seemed pointless to try to hack
the source to pass Aztec as (I don't have MASM) just to do something EDPART
does better.
> Partition information:
> Drive: ST225, Controller: WD1002-WX2 without autoconfig
> 	17 sectors/track, 4 heads, 611 cylinders (512 bytes/sector)
> Partition 1:	cylinders 000-150, 10267 sectors (DOS, bootable)
> Partition 2:	cylinders 151-301, 10268 sectors, 5134 blocks
> Partition 3:	cylinders 302-452, 10268 sectors, 5134 blocks
> Partition 4:	cylinders 453-609, 10676 sectors, 5338 blocks
> 	(cylinder 610 is reserved for head parking)
I have the exact same type of controller and drive, except that mine's 
autoconfig.  As far as I know the only difference between the two types of
controller is the ROM; does anyone else know for certain?
I've been running for about a week now with the new driver and have one
DOS and three MINIX partitions:
   Partition 1:  cyl. 0-480 (32691 sectors) DOS (4) bootable
   Partition 2:  cyl  481-586 (7208 sect)   DOS (4) extension |
   Partition 3:       587-598 (816 sectors) ""                |--for MINIX
   Partition 4:       599-610 (816 sectors) ""                |
Reads and writes on all 3 MINIX partitions work fine.  I've done several
kernel recompilations from the hard disk.
(You may wonder why the DOS partition is so big.  Apparently if the DOS
partition is smaller than this FORMAT won't format it in DOS 3.0 style but
will instead use 2.0 style with its larger cluster sizes.  Does anyone know
why DOS does this?)
> I am at a loss for this.  Presumably the driver works, I know the drive is
> working right (I reformatted before doing this); has anyone used EDPART.EXE
> yet?  Is it perhaps broken?
As I mentioned before, I used EDPART to set up the MINIX partitions.  It
seems to work except that for the DOS partition you should use FDISK.  FDISK
when creating a DOS partition does some DOS-specific initialization that
EDPART doesn't do.  Format tends to get confused if you change the DOS
part. with EDPART instead of FDISK.  What I did was to run FDISK, delete
all partitions, and create a DOS partition of the requisite size.  Then I
used EDPART to make the MINIX partitions.  (Needless to say, it took a
good bit of trial and error to hit upon this sequence.)
  Apart from the above-mentioned problem on DOS partitions, EDPART works
fine.  Just remember that when it says "You must REBOOT before partition
changes are recognized" it means it.  (I found this out when I failed to
do so and made a 16M DOS partition that got formatted as if it were still
the whole disk.  Strangely enough, CHKDSK never noticed that there more
blocks accounted for in the FAT than in the partition info.)
  This may be an obvious suggestion, but have you tried putting print
statements in the driver (at the end of init_params) to print out just what
MINIX thinks the partitions are when it boots?
--------------------------------------------------------------------------
Richard Todd
USSnail:820 Annie Court,Norman OK 73069
UUCP: {allegra!cbosgd|ihnp4}!okstate!uokmax!rmtodd