Copyright (c) 1980 Regents of the University of California.
All rights reserved. The Berkeley software License Agreement
specifies the terms and conditions for redistribution.

@(#)up.4 4.1 (Berkeley) 05/15/85

C 4 up - unibus storage module controller/drives controller sc0 at uba? csr 0176700 vector upintr

disk up0 at sc0 drive 0

This is a generic UNIBUS storage module disk driver. It is specifically designed to work with the Emulex SC-21 controller and Ampex or CDC 300M or Fujitsu 160M drives. It can be easily adapted to other drives and controllers (although bootstrapping will not necessarily be directly possible.)

Files with minor device numbers 0 through 7 refer to various portions of drive 0; minor devices 8 through 15 refer to drive 1, etc. The standard device names begin with ``up'' followed by the drive number and then a letter a-h for partitions 0-7 respectively. The character ? stands here for a drive number in the range 0-7.

The origin and size of the pseudo-disks on each drive are as follows:

AMPEX/CDC 300M drive partitions:
 disk start length cyl
 up?a 0 15884 0-26
 up?b 16416 33440 27-81
 up?c 0 495520 0-814
 up?d 341696 15884 562-588
 up?e 358112 55936 589-680
 up?f 414048 81472 681-814
 up?g 341696 153824 562-814
 up?h 49856 291346 82-561

FUJITSU 160M drive partitions:
 disk start length cyl
 up?a 0 15884 0-49
 up?b 16000 33440 50-154
 up?c 0 263360 0-822
 up?h 49600 213760 155-822

It is unwise for all of these files to be present in one installation, since there is overlap in addresses and protection becomes a sticky matter. The up?a partition is normally used for the root file system, the up?b partition as a paging area, and the up?c partition for pack-pack copying (it maps the entire disk). On 160M drives the up?h partition maps the rest of the pack. On 300M drives both up?g and up?h are used to map the remaining cylinders.

The block files access the disk via the system's normal buffering mechanism and may be read and written without regard to physical disk records. There is also a `raw' interface which provides for direct transmission between the disk and the user's read or write buffer. A single read or write call results in exactly one I/O operation and therefore raw I/O is considerably more efficient when many words are transmitted. The names of the raw files conventionally begin with an extra `r.'

In raw I/O counts should be a multiple of 512 bytes (a disk sector). Likewise seek calls should specify a multiple of 512 bytes.

/dev/up[0-7][a-h] block files

/dev/rup[0-7][a-h] raw files

hk(4), hp(4) up%d%c: hard error sn%d cs2=%b er1=%b er2=%b. An unrecoverable error occured during transfer of the specified sector of the specified disk partition. The contents of the cs2, er1 and er2 registers are printed in octal and symbolically with bits decoded. The error was either unrecoverable, or a large number of retry attempts (including offset positioning and drive recalibration) could not recover the error.

up%d: write locked. The write protect switch was set on the drive when a write was attempted. The write operation is not recoverable.

up%d: not ready. The drive was spun down or off line when it was accessed. The i/o operation is not recoverable.

up%d: not ready (came back!). The drive was not ready, but after printing the message about being not ready (which takes a fraction of a second) was ready. The operation is recovered if no further errors occur.

up%d%c: soft ecc sn%d. A recoverable ECC error occurred on the specified sector of the specified disk partition. This happens normally a few times a week. If it happens more frequently than this the sectors where the errors are occuring should be checked to see if certain cylinders on the pack, spots on the carriage of the drive or heads are indicated.

sc%d: lost interrupt. A timer watching the controller detecting no interrupt for an extended period while an operation was outstanding. This indicates a hardware or software failure. There is currently a hardware/software problem with spinning down drives while they are being accessed which causes this error to occur. The error causes a UNIBUS reset, and retry of the pending operations. If the controller continues to lose interrupts, this error will recur a few seconds later.

In raw I/O read and write (2) truncate file offsets to 512-byte block boundaries, and write scribbles on the tail of incomplete blocks. Thus, in programs that are likely to access raw devices, read, write and lseek (2) should always deal in 512-byte multiples.

DEC-standard error logging and bad block forwarding should be supported; the code to do this could be easily incorporated from the hp (4) driver. All that would be needed then would be a stand-alone formatting program to detect the bad sectors, format the disk so that the sectors were marked bad and initialize the bad sector files.

A program to analyze the logged error information (even in its present reduced form) is needed.

The partition tables for the file systems should be read off of each pack, as they are never quite what any single installation would prefer, and this would make packs more portable.