Digital Microsystems HiNet Protocols
DIGITAL MICROSYSTEMS, INC.
HiNet Protocols for BIOS 2.249
Version 1 .1
Notice:
Digital Microsystems, Inc. reserves the right to make
improvements to the products described in this manual at any time
without notice.
All rights reserved. No part of this publication may be
reproduced without the prior written permission of Digital
Microsystems, Inc. Please call: (415) 532-3686.
Release: 06/21/84
Digital Microsystgns
HiNet Protocols
HiNet Protocols for BIOS 2.249
This document can be used by anyone desiring to write
programs or design systems which interface with HiNet. HiNet is
a CP/M and MS-DOS compatible local microcomputer network de-
veloped by Digital Microsystems. The features and commands of
HiNet are presented in the DMS-3/4 HiNet System Manual. It is
assumed that the reader of this document is familiar with the
system manual, and well acquainted with HiNet.
Section
Number
Section
Title
Section
Contents:
1
Network Transmission
(page 2)
A description of the for-
mat of each and every
type of network transmis-
sion.
HiNet Master
(page 4)
Overview of the functions
performed by the network
master .
Login Procedure
(page 8)
Description of the login
procedure .
HiNet Comnands
(page 20)
HiNet BIOS
(page 49)
A complete description of
each network transaction.
The BIOS interface proto-
cols to be used by an
application program (such
as WHO).
SENDNET & RBCNET
(page 54)
The SENDNET and RECNET
routines. Listings are
discussed along with a
discussion of the pecu-
liaritiesof programming
the Zilog SIO chip for
SDLC communications.
Release: 06/21/84
Section 1
Network Transmission Format
Section 1 : Network Transmission Format
Each transmission on the network is done in Synchronous Data
Link Control (SDLC) format. SDLC was introduced by IBM for
computer- to-computer communication. It was chosen for HiNet
primarily because the widely available Zilog SIO chip implements
most of the details of SDLC transmission and reception. The
HiNet system automatically programs the SIO and DMA chips to send
or receive blocks of data appropriately. Each SDLC transmission
has the following format:
Flag
byte
User
number
Data bytes CRC bytes
(1 to 1024 bytes) (2 bytes)
Flag
byte
Field Description
Flag byte A flag byte is the bit sequence '01111110'. At
least two flag bytes surround each transmission.
The SDLC standard requires a minimum of one flag
byte before and after each transmission. However,
HiNet forces several flag bytes at both ends
because it is suspected that the SIO chip has a
bug which causes it to miss a flag occasionally.
User number Each station is assigned a unique identification
number, a user number, when it logs in. Each and
every transmission to a station must include its
user number. The master is always assigned user
number 0, while all other stations are assigned
numbers from 1 to 63. User numbers 251 thru 254
are reserved for special purposes, which are des-
cribed in section 2.
Data bytes One or more bytes can be transmitted in the data
portion of an SDLC transmission. In HiNet, the
data bytes may specify a command, a response, or
data read from or to be written to the master
disk.
CRC bytes
Each transmission is terminated by 16 bits of
error-check information. These bits are computed
when data is transmitted and are re-computed when
data is received. If an error occurs in the
middle of a transmission, the usual result is a
detectable CRC error. HiNet will retry any net-
work transaction which has a CRC error.
Release: 06/21/84
Section 1 Network Transmission Format
Whenever it observes five consecutive ones in the data
stream, the SIO chip inserts a zero bit automatically. These
extraneous zero bits are removed by the receiving SIO chip. This
zero-insertion method allows the chip to recognize flags, and
thus to identify the beginning and the end of each data
transmission.
Release: 06/21/84
Section 2
HiNet Master Functions
Section 2: HiNet Master Fonctions
On the master station, the basic control loop is as follows:
1. The master process is invoked each clock tick (usually 62
hertz). When the local user engages a private floppy disk opera-
tion, the master process is delayed until the next clock tick.
This is necessary because the DMA chip is shared by floppy and
network operations.
2. The master polls each active user after invocation. Active
users can respond with one of the commands listed below. All
users that have acknowledged the Master's previous poll are
active. A poll is a one byte command (50h). All network trans-
missions are done in SDLC format, so the poll is actually pre-
ceded by the one-byte destination user number.
Description of
HiNet Conmand
Command
Acknowledge 41 h
Get who table lOh
Read 128 bytes llh
Read 1024 bytes ... 15h
Write 128 bytes ... 12h
Start spool file .. 14h
Spool 128 bytes ... ICh
End spool file .... 1 6h
Assign partition .. 17h
Hog the network ... 1 8h
Poll Prime 55h
Lock record 1 9h
Unlock record 1 Ah
Clear all locks ... iBh
Get HD status IDh
Get date time lEh
Login 13h
Instant logout .... IFh
Write Modes 20h
Network info 21 h
Comntiand
length
1 byte
1 byte
8 bytes
8 bytes
8 bytes
2 bytes
8 bytes
2 bytes
15 bytes
1 byte
1 byte
15 bytes
1 5 bytes
1 byte
1 byte
1 byte
20 bytes
2 bytes
6 bytes
1 byte
Additional command
parameters
.. dtn,src,dsk,trk,sec,vli
., dtn,src,dsk,trk,sec,vli
.. dtn,src,dsk,trk,sec,vli
.. sid
.. dtn,src,dsk,trk,sec,vli
.. sid
.. nam,psw
.. len,lck
.. len,lck
.. usr,psw,ser#,prod
.. src
.. wmc,vli,dsk,val,usr
Release: 06/21/84
Section 2 HiNet Master Functions
Coomand Parameter
Parameters Length
dtn = destination station nuniber (always 0) 1 byte
src = source station number (same as user number) 1 byte
dsk = partition number (0-63) 1 byte
trk = track number (0-511) 2 bytes
sec = sector number (1-128) 1 byte
vli = volume number (0-3 ) 1 byte
nam = partition name 8 bytes
usr = user name 8 bytes
psw = password 6 bytes
len = length of lock string (1-13) 1 byte
Ick = lock string 13 bytes
wmc = write mode command (grt/rel/frc/qry) 1 byte
val = write mode force value or logical user number ... 1 byte
usr = write mode physical user number 1 byte
sid = spool job id number 1 byte
Each of these commands is described in detail in section 4.
3. If a user responds to a poll with an "acknowledge" command,
then no further interaction with the user v;ill be contemplated
until the master process re-awakes on the next clock tick. All
other commands require an interchange of information between the
master and the user station, as described in section 4.
4. A user station that fails to respond to 256 consecutive polls
in the following manner will get logged out: the user is polled
at the normal polling rate 160 times. The station then becomes a
'slow user' and is polled at 1/12 the normal polling rate for an
additional 96 times. A user that responds to any of these polls
regains normal status; otherwise, the user is logged out. When a
user is logged out his user number and all of his locks are
released. If a spool file is being created, it is erased. If the
user owns any partitions, they are released.
5. After polling all active users, the master checks the local
user for pending requests. If a request is pending, the local
user's network command byte will be non-zero. Pending requests
are processed, and their command bytes are set to zero, signaling
command completion.
6. The Z80 network station bootstrap code is transmitted period-
ically (once per polling loop and about once per second) to
pseudo-user 254. The PROM at each Z80 station has been programmed
to receive 380h bytes addressed to 254's user number so that it
can boot from the network. The bootstrap code is loaded into
memory at location 9000h, and executed. This boot code displays
the "HiNet 2.2xx" message, waits for a poll of pseudo-user 253,
and attempts to log in by PROM serial number. The 8086/8088
Release: 06/21/84
Section 2 HiNet Master Functions
network station PROM contains sufficient code to attempt the
login by PROM serial number directly.
7. The master polls pseudo-user 253 periodically (once per
polling loop and about once per second). Any station that wants
to connect to HiNet must first log in by responding to a poll of
user 253 with a HiNet user name, password, binary serial number,
and product type. The master consults the Machine table. Product
Type table, and User table on the hard disk for a matching name,
password, and product type. The master accepts the login request
unconditionally and responds with a unique user number, the login
time, and the binary serial number. Then it will immediately send
boot phase 2, a loader for the BIOS or Login Please program.
What is loaded depends on whether the name/password was found in
the User table and the network station's product type.
8. The master polls pseudo-user 252 once per master polling
loop. A mimicking system (if any) must respond to these polls to
come on-line and remain on-line.
9. The master checks regularly (once every polling loop and also
about once per second) whether the spool print buffer is empty
and needs to be refilled. If so, the next sector of the printing
spool file is read, and printing is restarted. If not, the
spooler checks whether a new spool file should be opened and
printed.
Release: 06/21/84
Section 2
HiNet Master Functions
The following tables and buffers are maintained in the
master:
Location
see Net info,
currently at
OFF Oh
Name of table
Spool file table
Description
Each spool file is described
by a 1 6 byte entry. See the
"Get who table" command in
section 4 for more informa-
tion.
see Net info,
currently
below spool
table
Lock table
see Net info. Who
currently
below lock
table
Each lock string is
described by a 1 6 byte
entry. The first byte is
the number of the user who
created the lock. If the
entry is not in use, the
first byte is OFFh. The
next byte is the lock string
length. The next 1 3 bytes
are the lock string. The
last byte is not used.
Each user is described by a
1 6 byte entry. See the "Get
who table" command in sec-
tion 4 for more information.
40 Oh bytes
below who
table
optional ,
lOOh bytes
below hard
disk buffer
Master Buffer
Floppy write buffer
General buffer for Master's
use. Not all hard disk I/O
operations use this buffer.
This buffer is used for all
double-density floppy write
operations.
optional ,
lOOh bytes
below floppy
write buffer
Floppy read buffer
This buffer is used for all
double-density floppy read
operations.
These tables can be manipulated while the network is running
by using DDT or ZDTI. However, extreme caution should be taken.
Release: 06/21/84
Section 3 Login Procedure
Section 3: Login E»rocedure
The entire login procedure, from a station's point of view,
is outlined below.
1. The Z80 PROM code programs the SIO-1 and the DMA chip to
receive 380h bytes addressed to user 254. When this code is
received, it is loaded into memory at location 9000h, and the
PROM executes a "JMP 9000h" instruction. The code sent is known
as Boot Phase 1. The 8086/8088 PROM contains Boot Phase 1
internally and therefore ignores this step.
2. An automatic login is attempted as follows: the eight
character name field contains the eight ASCII hexadecimal digit
machine serial number; a six-character password of blanks; binary
machine serial number, in four bytes; product type, in one byte.
The SIO chip is then programmed to receive a poll from user 253.
When a poll is received (command byte 50h) , the station will
respond with a command byte of LoginNet (1 3h) , followed by the
automatic login data. The master will normally respond with a
LogAck (4Ch) message. The LogAck response byte will be followed
by: a unique user number and login time (i.e., the 7 bytes stored
in locations 40h-46h on the Z80) and the machine's binary serial
number (in four bytes).
The station will verify that the LogAck (4Ch) response was
followed by the station's 4-byte serial number; this provides an
additional level of error checking beyond that which the SIO
provides in hardware.
The master will send a LogNack (4Eh) to pseudo-user 253 if
no login request is received within approximately 1 mSec. Note,
if two stations attempt to log in simultaneously, it is likely
that CRC errors will be generated and each station will receive a
response of LogNack (4Eh). If either a flawed network transmis-
sion or a LogNack (4Eh) is received, the Z80 station will use the
refresh register to pause (from 400ms to 16 seconds), and will
then attempt another automatic login. The 8086/8088 will use an
algorithm based on its serial number to pause before retry.
If the ASCII serial number and password are not in the user
table, the master will still respond with a LogAck (4Ch) .
However, when Boot Phase 2 (described in step 3 below) is sent, a
Login Please program will be selected to run in the station. The
user will be prompted for a login name and password, and the
login procedure thru pseudo-user 253 will be repeated as above.
If no Boot Phase 2 can be found for the given product type,
or any crucial table lookup failed, e.g., a Login Please program
could not be found, the master will respond with a LogDeny (44h)
and the same parameter list as LogAck.
Release: 06/21/84
Section 3 Login Procedure
3. Once a login request has been accepted, the master will
follow the "login accepted" transmission with an immediate trans-
mission (addressed to the number of the user who just logged in)
of additional bootstrap code. Boot Phase 2. This procedure is
necessary because the entire bootstrap code will not fit within
380h bytes: it must be transmitted in 2 parts. The second part
of the bootstrap contains a loader (Boot Phase 2) and a load
list. The loader is sent in one or more 1024 byte transmissions,
depending on the value of the first byte sent. The load list
contains the disk addresses, lengths, and transfer addresses of
the BIOS/BDOS/OCP or Login Please program which the master has
selected based on the Product Type and User name. The loader
reads the station's BIOS or other appropriate program from the
master hard disk. Note, once a user has been granted a unique
user number, that user number is used for all following network
transactions. Also note, the user is allowed to initiate a
network transaction only after receiving a poll. If the loader
reads in a Login Please program instead of a BIOS, that program
will do an Instant Logout (20h) after receiving the first poll,
prompt the user to enter a name and password, and repeat the
login sequence similar to automatic login above.
Release: 06/21/84
Section 3
Login Procedure
Hard Disk Control Area Layout
(Partition 0)
Logical Address
track 0, sectors 01 -IF
track 0, sectors 20-28
track 0, sectors 29-38
track 0, sectors 39-78
track 0, sectors 79-80
Contents
Controller Program
reserved for expansion of
controller program
HiNet User Name Table
Up to 128 16-bYte entries:
8 bytes: user name or
station serial no,
6 bytes: password
1 byte: OS code
1 byte: flags (incl
big/small
request)
HiNet User Configuration Table
Up to 128 64-byte entries:
8 bytes: default A drive
8 bytes: default B drive
8 bytes: default C drive
8 bytes: default D drive
1 byte: length of
typeahead
31 bytes: typeahead buffer
Disk Allocation Table
Up to 64 16-byte entries:
1 byte: size (0-8)
8 bytes: partition name
6 bytes: password
1 byte: control byte
Release: 06/21/84
10
Section 3
Login Procedure
track 1 , sectors 01 -08
track 1 , sectors 09-1 4
track 1 , sectors 1 5-1 6
track 1 , sector 1 7
track 1 , sector 1 8
track 1, sectors 19-20;
Bad Sector Table
Up to 64, 128, or 256 3-byte
entries depending on drive type:
1 byte : track
1 byte : head
1 byte: sector
Machine Table
Up to 128 12-byte entries:
4 bytes: Serial Number
1 byte: Product Number
6 bytes: Option Map
1 byte: lOBYTE
Write Mode Table
reserved
Password Table
Product Type Table
Up to 40 25-byte entries:
1 byte: Product Type
8 bytes:
8 bytes:
8 bytes:
Boot Phase 2
program name
Login Please
program name
OS Menu program
name
track 1, sectors 21-80:
OS Table
Up to 128 96-byte entries:
1 byte:
1 6 bytes :
6 bytes:
64 bytes:
9 bytes:
OS number
Product Map
Option Map
Load List (8 names
of 8 bytes)
— reserved —
Release: 06/21/84
11
Section 3
Login Procedure
track 2, sectors 01-02
track 2, sectors 03-08
track 2, sectors 09-20
track 2, sectors 21-80
Cold Boot Loader
reserved for use of Cold Boot
Loader
Systen Directory
Up to 128 24-byte entries:
8 bytes: File Name
5 bytes: Disk Address
2 bytes: Length (128-byte
records)
4 bytes: Load Address
2 bytes: Execution Address
Offset
1 byte: Program/Data flag
2 bytes: — reserved —
— reserved —
Remainder of partition allocated according to contents of the
Systan:! Directory.
Release: 06/21/84
12
Section 3 Login Procedure
BOOT PHASE 2 program
Boot Phase 2 is the loader for the OS to follow.
Its first byte will be the number of 1 024-byte blocks the Boot
Phase 2 program occupies, enabling the receiving station to set
up for another network transfer if necessary.
The next three bytes enable the host to jump to the beginning of
the Boot Phase 2 code. In a Z80 this will simply be a jump
instruction. In an 8086/8088 the first of these bytes will be
null and the next two will be the offset from the beginning of
Boot Phase 2 to the beginning of the executable code.
Next cones a data block, always in the following format:
Default partition assignments
(four names of eight characters) 32 bytes
lOBYTE 1 byte
Default type-ahead buffer and pointer 32 bytes
Honor Flag (see description below) 1 byte
Load List
(up to eight partial directory entries
each 16 bytes, of the form:
Disk Address (v,p,t,s) 5 bytes
length (128-byte sects) 2 bytes
RAM Load Address 4 bytes
Relative Start Address 2 bytes
Reserved 3 bytes)
total: 128 bytes
Product Type 1 byte
total length of data block 1 95 bytes
This data block is initialized by the Master before Boot Phase 2
is sent out. A description of each field follows:
Default partition assignments:
These are the assignments specified in the User Configuration
Table entry (created by USERS), if a BIOS is to be loaded.
If a Login Please or OS Menu is to be loaded, these fields are
undefined.
lOBYTE:
This comes from the Machine Table entry for this machine. If
there is no Machine. Table entry, the lOBYTE will be 00 and the
Boot Phase 2 program should attempt to determine the console type
and set this field correctly. If this is done then a BIOS loaded
by this Boot Phase 2 will be able to run in at least a minimal
fashion, enabling the user to run MACHINE and put the machine in
the Machine Table.
Release: 06/21/84 13
Section 3 Login Procedure
Default type-ahead buffer:
As specified in the User Configuration Table (set up by
USERS), if the user name/password is found. Otherwise, undefined.
Honor Flag:
This flag is set to show the extent to which the login re-
quest was honored, i.e. what sort of program will be loaded next.
The lower nibble will be:
- An OS exactly matching the request will be sent, i.e.
its Option Map will match the machine's (unless a
high-TPA BIOS was requested.)
1 - A default OS will be sent, of the general type as
the requested one but with a different option map.
Implies that there is no OS available with the
exact option configuration of the machine.
2 - A "login please" program will be sent. Implies
that the name/psw were not found in the User
Table .
3 - An "OS Menu" program will be sent. Implies that
no OS of the requested type could be found that
would stand a chance of running on the machine
logging in. In the initial release this will
probably be a message rather than a real menu.
4 - Total failure, no program will be loaded.
The high bit of the honor flag is used to show whether the
machine is in the Machine Table, viz.:
- Normal (the machine is in the Machine Table.)
1 - Not found. Without a Machine Table entry. Boot
Phase 2 will have to try to figure out the
correct lOBYTE for this machine, probably by
examining the PROM.
Depending on the value of the Honor Flag, Boot Phase 2 may
print a message (using PROM I/O routines) notifying the user of
an unusual situation (e.g. a default OS or a Machine tTble lookup
failure.)
Load List:
This list consists of entries from the System Directory, with
the eight-byte names removed, making each entry sixteen bytes
long. Each entry describes a file to be read from the net,
including information as to where it is to be loaded and where to
begin executing it. If fewer than eight files are specified
(actually two will be a more common number), the length field in
the entry after the last one used is set to zero by the master
(in fact, the block is filled with zeros on the end, but these
might be legal values in other fields.)
Release: 06/21/84 14
Section 3 • Login Procedure
Product Type:
Included in Boot Phase 2 because Z80 machines do not know
their own product type and so must find out from the Master.
Thus there are at least 1 99 bytes in Boot Phase 2 before the real
executable code.
Release: 06/21/84 15
Section 3 Login Procedure
"LOGIN PLEASE" program
This program will be loaded in lieu of an OS if the name/psw in
the login request do not match any entry in the User Table.
(This will often be true on auto-login by serial number.)
Its first action will be to request an "instant logout" of the
current user number. The program will then ask the user for a
name and password. After appending the binary serial number,
and product type, the program will wait for a poll of the login
pseudo-user (253), issue a login request, and prepare for the
log-ack.
The Master's response will be handled the same by this program as
it would be in the above boot procedure; i.e., the serial number
in the response will be matched against our own, and the login
request will be retried if the serial numbers don't match. If
they do match, and the response is an ack, the program notes the
new HiNet user number and waits for a new Boot Phase 2 to be sent
to that user number. It then executes Boot Phase 2. If the
serial number matches but the response is a nack, this is a fatal
error. A diagnostic message will be sent to the screen.
The login program will have access to the PROM I/O facilities, to
the information passed in Boot Phase 2, and to the HiNet user
number.
Release: 06/21/84 16
Section 3 Login Procedure
"OS MENU" program
This program loads in lieu of an OS. It uses PROM I/O facilities
and has access to the information passed in the Boot Phase 2 code
and to the Hi Net user number.
The program reads the OS table from the master and presents to
the user a menu of the OS's capable of running on the machine
logging in. (Presenting the option map in a reasonable way may
be difficult.) When the user selects one, the program loads it
— probably by reinitializing the still-resident Boot Phase 2
with the address of the selected OS, and re-running Boot Phase 2.
Since appearance of the "OS MENU" program implies that no system
of the type specified in the User Option Table could be found for
the machine, it is unlikely that the default partitions will be
suitable for the OS eventually selected. Therefore the re-
initialization of Boot Phase 2 must include new default par-
titions as specified by the user from the console. Program must
check that the requested partitions can be used under the re-
quested OS. Any TPA utility version of this program will also
have to ask for the default partitions, for the same reason.
Note that this program is required to know the location of the OS
table in the master.
Until completion of the OS menu program, the system would benefit
from a skeletal program that tells the user why the login request
could not be honored.
OS numbers
not to be used - can be returned if search fails, etc
1 CP/M
11 = CP/M 2
12 = CP/M 86
13 = HiDOS
14 through IF = other CP/M compatible OS's
2 MS-DOS
21 = MS-DOS vers. 2.11
22 = MS-DOS vers. 2.x
23 = MS-DOS vers. 3.x
24 through 2F = other MS-DOS canpatible OS's
3 Reserved
4 through F not yet assigned
Release: 06/21/84 17
Section 3
Login Procedure
Product Numbers
The highest bit of the Product Number specifies the console type
(0 for serial, 1 for parallel). This leaves 128 possible product
numbers which specify the type of CPU board as follows:
Not to be used
1
ZSBC-3 CPU's:
DMS-3 "DSC-3"
DMS-3/A25 "Smart ADDS"
DMS-3/4004
DMS-3/4008
DMS-3/1 01
DMS-3/1 02
DMS-3/1 03
DMS-3/B
DMS-3/F "Fox"
DMS-501 /I 5
DMS-5080
2
DMS-4 "DSC-4 "
3
DMS-1 280
4
DMS-3/C "Killer Bee"
5
DMS-5086
and HNS-86
6
reserved
7
DMS-81 6
8
PC Adapter
(Product numbers 9 through 127 not yet assigned)
Release: 06/21/84
18
Section 3 Login Procedure
Optional Device Drivers
This is a bit-map with the exception of the block describing the
number of logical drives supported. An exact match with the
Machine Table is required. Since it is possible to configure an
arbitrary number of logical drives (up to 16), the only bit-
mapped way of doing it would use 16 bits or one- third of the bit
map.
N.B. There is no way of knowing, a priori, whether a driver
named "custom printer" will run on a given machine.
bit device driver
8- inch floppies (SD and DMS DD)
1 5- inch Fox- floppies
2 5- inch IBM- format floppies
3 8- inch HD with DMS controller
4 5- inch HD with Xebec controller
5 5- inch HD with Adaptec controller
6 Port type- ahead (console)
7 Port polled ( " )
8 Port 2 polled (printer)
9 Port 3 type-ahead (aux ccmn.)
1 Port 3 polled ( " " )
11 Parallel Port 1 (console)
12 Parallel Port 1 (printer)
13 Parallel Port 2 (Fox printer)
14 Console/Printer Mux (ADDS, 1280, 5000)
1 5 Spooler
16 Net Buffer (Ik)
17 Real-Time Clock
18 Front-Panel Interri:5)t
19 Number of logical drives (bit 0)
20 Number of logical drives (bit 1 )
21 Number of logical drives (bit 2)
22 Number of logical drives (bit 3)
23 Manory-mapped console
Release: 06/21/84 19
Section 4 HiNet Commands
Section 4: HiNet Ccmnaiids
Each of the HiNet command protocols is described in this
section. A diagram showing the exchange of information between
the master station and the user station is included for each
command. The direction of information flow is shown by the
arrows, and the format of the information is given in parentheses
above the arrow. The approximate time required to send the
information is given below the arrow. The time between exchanges
of information (i.e., turnaround) is usually about 0.5ms, exclu-
ding disk access time. A disk access may require anywhere from
2ms to 150ms to complete. For example, the following diagram
shows the poll-acknowledgment sequence:
poll (50h)
master > user
station 0.16ms station
acknowledge (41 h)
<
. 1 6ms
Total time (including turnaround): 0.82ms
A poll is a one-byte message sent from the master station to
user stations around the network; an acknowledgment is a one-byte
message sent to the master from the network user stations. The
BIOS in a HiNet station normally answers all polls in this
fashion; this acknowledgment is considered to be one of the
network commands — its function is to tell the master that the
station is active, but does not have any request for the master.
The basic polling rate is approximately 62 times a second, with
stations which read consecutive records from the master hard disk
being polled continuously until they cease to make consecutive
read requests. Network loading and the preferential polling given
to stations making consecutive read requests will affect the
frequency with which any given station is polled.
Release: 06/21/84 20
Section 4 HiNet Commands
Error Detection and Recovery
When the network master receives either a command or
response that is illegal or unrecognizable it will print the
message shown below. But note, this applies only when the error
message display flag is true.
"*** User WW xxxxxxxx yyyy Net error at zzzz"
"MASTbuf : Mst Usr To Frm Dsk Track Sec Vol DmaAdr"
" aa bb cc dd ee ff gg hh ii jj kk"
WW = user number in hexadecimal
xxxxxxxx = user name
yyyy = error type [ LATE, CRC, OVR, SYNC ]
zzzz = error address in hexadecimal
aa - kk = command specific parameters in hexadecimal
The master will return immediately to the main polling
loop, regardless of the error message display flag state; the
user should be polled during the next invocation of the master.
The user station functions similarly, except:
1 .) the user name is not printed;
2.) "NET buf:" is displayed instead of "MASTbuf:"; and
3.) whatever function the station was executing is retried
automatically.
THESE MESSAGES INDICATE THAT AUTOMATICALLY CORRECTED
TRANSIENT ERRORS OCCURRED - NETWORK INTEGRITY IS GUARANTEED. NO
USER INTERVENTION IS REQUIRED.
Since MASTbuf is a general purpose buffer, the meaning of
the command-specific hexadecimal parameters can vary greatly. In
general, Mst represents a command that the master sends to a
station, for example, a poll (50). Usr represents the command
the station sends back to the master, for example, a read or
write (11, 12 or 15). To and Frm generally represent user
numbers that messages are sent to or received from. Dsk, Track,
Sec, Vol, are usually meaningful for read and write, and DmaAdr
may indicate an address in the Master's memory. Not every field
may be meaningful even though it is displayed. Error types have
the following meaning:
LATE — Master did not receive an expected trailing
SDLC flag byte, which indicates that the
station did not complete the transaction.
CRC — Master received data with transmission errors
OVR — Master received more data than was expected
SYNC — Master received unexpected data or comnand
Release: 06/21/84 21
Section 4 HiNet Commands
Mimicking
Under HiNet, one can connect a second hard disk system to
the network cable and duplicate all master disk writes on that
system. The second system is called the "mimicker." The mimicker
is brought on-line by answering the master's poll of pseudo-user
252. This pseudo-user is polled approximately 3 times per
second. The mimicker prints an 'M' on its console whenever it is
polled. The mimicker must answer each poll to remain on-line.
While the mimicker is on-line, all writes will be mimicked.
The following diagram shows the master-mimic polling
sequence :
poll (50h)
master > mimic
station 0.16ms station
(user 252)
acknowledge ( 41 h)
0.16ms
Total time (including turnaround): 0.8ms.
The other mimicking protocols are described under the "Write
128 bytes" and "Spool 128 bytes" commands.
Release: 06/21/84 22
Section 4 HiNet Commands
Login CGnamaiid
The login command is used to connect to the HiNet system.
The entire login procedure is described in sequence. The login
command sequence is summarized by the following diagram:
poll (50h)
master > user
station 0.16ms station
(user 253)
login (1 3h, name, psw,ser#, prod)
<
0.46ms
login-nack (4Eh) OR
login-deny (44h,usr,time,ser#) OR
login-ack (4Ch,usr,time,ser#)
>
. 35ms
bootstrap code (1 or more 1 024 byte groups)
> to new
1 6 . 53ms user number
Total time (including turnaround, excluding disk accesses for
machine, product, and user table lookups): 19.00ms
The first byte of the bootstrap code sent indicates the number of
1024 byte transmissions necessary for the entire bootstrap
reception.
Coomand Parameter
Parameters Length
nam = user name 8 bytes
psw = password 6 bytes
usr = user number 1 byte
time= login time ( t ic ks, sec ,min,hr,mth, day, year) 7 bytes
ser#= binary serial number 4 bytes
prod= machine product number 1 byte
Release: 06/21/84 23
Section 4 HiNet Commands
Get Who and Spool Tables
The who command can be used by any station to determine who
is currently logged into HiNet, and who has active spool files.
poll (50h)
master > user
station 0.16ms station
who (lOh)
<
0.16ms
who table (513 or 1025 bytes)
>
8.34 or 1 6.53ms
spool table (256 bytes)
>
4.24ms
Total time (including turnaround): 14.40ms or 22.59ms
The who table has either 32 or 64 entries, each 16 bytes
long. The first byte of the who table sent indicates the number
of users. This value is also returned by Netlnfo. Each entry
corresponds to a single user. The first entry describes user
(the master user); the remaining entries describe users 1 to 31
or 63. The Netlnfo protocol outlined later on in this document
may also be used to determine the actual number of users. Each
entry contains the following information:
byte = 00 if entry not in use
= OFFh if user is logged in
= other if user is logging out (counting
down one for each missed poll)
bytes 1-8 = user name
bytes 9-11 = login time (sees, mins, hrs)
bytes 12-14 = time of most recent HiNet transaction
byte 15 = most recent HiNet command byte
Release: 06/21/84 24
Section 4 HiNet Commands
The spool table has 16 entries (the value currently returned
by Netlnfo) , each 16 bytes long. Each entry corresponds to a
single spool file. Each entry contains the following information:
byte
= 00
if starting
01
if spooling
02
if ready to print
03
if printing
04
if finished
05
if waiting
0E5h
if entry not in use
byte ^
= user
number
byte 2
= spool job id
bytes 3-
-4
= time of most recent spool request
bytes 5-
-6
= current spool track
byte 7
= current spool sector
bytes 8-
-15
= user
name
Release: 06/21/84 25
Section 4 HiNet Coinmands
Read 128 Bytes
This command allows a station to read 128 bytes from a
specified partition, track, and sector. The standard HiNet
station BIOS does not use this command — the "Read 1024 bytes"
command is used instead, because it is much more efficient.
The standard station BIOS will process the readl 28 command as a
1024 byte read and will return the appropriate 128 byte data
block to the application. Those applications which read or write
to a multi-user partition need to guarantee sometimes that data
is current and not outdated by other user modifications;
therefore, a DMS-specific BIOS jump table entry is provided which
allows an application to force the BIOS to throw away the Ik net
buffer contents and perform a network transfer.
The BIOS may be assembled to allow true 128 byte network
reads; however, in so doing, one sacrifices the ability to
perform 1 024 byte network reads.
poll (50h)
master > user
station 0.16ms station
read ( 1 1 h ,d tn , src ,dsk , trk , sec , vli)
<
0.27ms
data (128 bytes)
: >
2.19ms
data-ack (44h)
<
. 1 6ms
Total time (incl. turnaround, excl. disk seek/read time): 4.28ms
Canmand Parameter
Parameters Length
dtn = destination station number (always 0) 1 byte
src = source station number (same as user number) 1 byte
dsk = partition number (0-63) 1 byte
trk = track number (0-51 1) 2 bytes
sec = sector number (1-128) 1 byte
vli = volume number (0-3 ) 1 byte
Release: 06/21/84 26
Section 4 HiNet Commands
Read 1024 Bytes
This command allows a station to read 1024 bytes starting at
a specified partition, track, and sector. The sector number
should be 1 , 9, 17, ... or 121 (8n + 1). These sector numbers
correspond to physical sector boundaries, i.e., each read
operation will read a single IK physical sector from the disk.
The standard HiNet station BIOS sets aside a IK buffer for all
read operations.
poll (50h)
master > user
station 0.16ms station
read (1 5h ,dtn , src ,dsk , trk , sec , vli)
<
0.27ms
data (1024 bytes)
>
1 6 . 53ms
data-ack (44h)
<
. 1 6ms
Total time (incl. turnaround, excl. disk seek/read): 18.62ms
(disk read time from 2 to 150ms)
Comnand Parameter
Parameters Length
dtn = destination station number (always 0) 1 byte
src = source station number (same as user number) 1 byte
dsk = parti tion number (0-63 ) 1 byte
trk = track number (0-51 1 ) 2 bytes
sec = sector number (1 ,9,17, .. .121 ) 1 byte
vli = volume number (0-3) 1 byte
Release: 06/21/84 27
Section 4 HiNet Commands
Write 128 Bytes
The write command is used to write 128 bytes to a specified
partition, track, and sector. If the mimicker is on-line, the
writes will also be mimicked.
poll (50h)
master > user
station 0.16ms station
write (1 2h,dtn,src,dsk,trk,sec,vli)
<
0.27ms
cmd-deny (4Fh) OR
msg-ack (4Dh)
>
. 1 6ms
data (128 bytes)
<
2.19ms
data-ack (44h)
>
0.16ms
Total time (including turnaround, excluding disk write): 4.94ms
(disk write occurs after data-ack)
If a mimicker is on-line, these protocols will also be used:
write (12h,dtn,src,dsk,trk,sec,vli)
master > mimic
station 0.27ms station
msg-ack (4Dh)
<
0.16ms
data (128 bytes)
>
2.16ms
data-ack (44h)
<
. 1 6ms
Total time (including turnaround, excluding disk write): 4.25ms
Release: 06/21/84 28
Section 4 HiNet Commands
Canmand Parameter
Parameters Length
dtn = destination station number (always 0) 1 byte
src = source station number (same as user number) 1 byte
dsk = partition number (0-63) 1 byte
trk = track number (0-51 1 ) 2 bytes
sec = sector number (1-128) 1 byte
vli = volume number (0-3 ) * 1 byte
* The high bit of the volume number is set automatically by the
station if the station or local user is running HiDos (to verify
that a shared partition is written to only by a HiDos OS and not
a CPM-86 OS).
A write may be denied for several reasons:
1. A non-HiDos write is attanpted to a shared partition.
2. The partition is marked read-only.
3. The partition is owned for writing by another user.
The write mode query function may be used to determine the source
of the error.
Release: 06/21/84 29
Section 4 HiNet Coinmands
Start Spool File
To start a spool file, a user station must execute the
following cotinand:
poll (50h)
master > user
station 0.16ms station
start-spool (14h,sid)
<
0.18ms
response (04h,dsk,trk, sec, vli, size)
. >
0.26ms
Total time (including turnaround): 1.6ms
The master returns:
1 . ) an AckSpStart (04h) ,
2.) the spool partition number (dsk) ,
3.) the track and sector at which spooling should start,
4.) a spool block size
If dsk = 0, then the spool request is denied. This will
happen if there is no PRTSPOOL partition, or if the PRTSPOOL
partition is full.
The station should maintain a spool job id byte whose high
nibble is a job number (initialized to zero when the station
boots, and incremented when a spool job ends), and whose low
nibble is the spool block number (initialized to zero when a
spool job starts and incremented when a spool block is filled).
The station should spool to consecutive sectors and tracks,
beginning at the track and sector supplied by the master. Spool-
ing should occur on sectors 3 to 1 28 on the first spool track,
and 1 to 128 on subsequent spool tracks. The number of tracks is
specified by the spool block size field. The station is respon-
sible for enforcing this limit; if an attempt is made to spool
beyond this limit, the space allocated to other spool files may
be overwritten. Rather than surfeit the spool file , issue a
"End spool file" command after writing the last sector, increment
the low nibble of the spool job id byte and then issue a "Start
spool file" command to resume spooling on a new spool block.
Release: 06/21/84 30
Section 4 HiNet Cornmands
Comnnaiid Parameter
Parameters Length
dsk = partition number (0-63) 1 byte
trk = track number (0-511) 2 bytes
sec = sector number (1-128) 1 byte
vli = volume number (0-3 ) 1 byte
sid = spool job id number 1 byte
size= spool block size (1,2,4,8,16) 1 byte
Release: 06/21/84 31
Section 4 HiNet Coinnands
Spool 128 Bytes
The spool command is used to write spool data to the
PRTSPOOL partition. The spool command is identical to the write
command, except that a different command byte is used.
The station should spool to consecutive sectors and tracks,
beginning at the track and sector supplied by the master. Spool-
ing should occur on sectors 3 to 1 28 on the first spool track,
and 1 to 1 28 on subsequent spool tracks. The number of tracks is
specified by the spool block size field. The station is respon-
sible for enforcing this limit; if an attempt is made to spool
beyond this limit, the space allocated to other spool blocks may
be overwritten. Rather than surfeit the spool file , issue a
"End spool file" command after writing the last sector, increment
the low nibble of the spool job id byte and then issue a "Start
spool file" command to resume spooling on a new' spool block.
poll (50h)
master > user
station 0.16ms station
spool (lCh,dtn,src,dsk,trk,sec,vli)
<
0.27ms
msg-ack (4Dh)
>
. 1 6ms
data (128 bytes)
<
2.19ms
data-ack (44h)
>
0.16ms
Total time (incl. turnaround, excl. disk write): 4.94ms
(disk write occurs after data-ack)
Release: 06/21/84 32
Section 4 HiNet Commands
If a mimicker is on-line, then these protocols will also be used:
spool (lCh,dtn,src,dsk,trk,sec,vli)
master > mimic
station 0.27ms station
msg-ack (4Dh)
<
. 1 6ms
data (128 bytes)
>
2.19ms
data-ack (44h)
<
. 1 6ms
Total time (incl. turnaround, excl. disk write): 4.28ms
Command Parameter
Parameters Length
dtn = destination station number (always 0) 1 byte
src = source station number (same as user number) 1 byte
dsk = partition number (0-63 ) 1 byte
trk = track number (0-51 1) 2 bytes
sec = sector number (1-128) 1 byte
vli = volume number (0-3)* 1 byte
* The high bit of the volume number is reserved, and will be set
by HiDos stations, however, a spool partition cannot be marked
HiDos shared.
Release: 06/21/84 33
Section 4 HiNet Commands
End a Spool File
Use this command to end a spool file, or swap to the next
spool block, making the current block available for printing.
The station should maintain a spool job id byte whose high
nibble is a job number (initialized to zero when the station
boots, and incremented when a spool job ends), and whose low
nibble is the spool block number (initialized to zero when a
spool job starts and incremented when a spool block is filled).
When the station reaches the end of a spool block (as
specified by the spool block size field of the start spool
protocol) , the station should issue an "End spool file" command
after writing the last sector, increment the low nibble of the
spool job id byte and then issue a "Start spool file" command to
resume spooling on a new spool file. The station is responsible
for enforcing the spool block size limit; if an attempt is made
to spool beyond this limit, the space allocated to other spool
blockss may be overwritten.
poll (50h)
master > user
station 0.16ms station
stop-spool (16h,sid)
<
0.18ms
msg-ack (4Dh)
>
. 1 6ms
Total time (including turnaround): 1.5ms
Caranand Parameter
Parameters Length
sid = spool j ob id number 1 byte
Release: 06/21/84 34
Section 4 HiNet Commands
Assign Partition
The assign command is used to determine the volume number,
partition number and size of a specified partition on the HiNet
shared disk. A station should not attempt to access a HiNet
partition (except partition 0) without first using the assign
command. The partition number should be used for all subsequent
access to the HiNet disk; the partition size should be used to
construct a disk parameter table required by the BDOS.
poll (50h)
master > user
station 0.16ms station
assign (1 7h,nam,psw)
<
. 38ms
response (siz,dsk,ctl,vli)
>
0.21ms
Total time (incl. turnaround, excl. alloc table lookup): 1.75ms
The master returns the partition's size, number, and control
byte. A size of OFFh indicates that the assignment was denied.
This will happen if the partition name and password did not match
any entry in the allocation table. To force the master to ignore
the password, use a password of all zeroes.
Gannand Parameter
Parameters Length
nam = parti tion name ' 8 bytes
psw = password 6 bytes
siz = partition size (1-6) 1 byte
dsk = partition number (0-63) 1 byte
ctl = control byte 1 byte
vli = volume number this parition is on 1 byte
Release: 06/21/84 35
Section 4
HiNet Commands
The contents of the disk parameter table for each possible
partition size is shown below:
CP/M 2.2 Disk Parameters
Sectors per track
Block shift, mask
Block count - 1
Directory count -
Directory blocks
Check vector size**
Cp sys tracks
Sectors per track
Blockshi f t ,mask
Block count - 1
Directory count -
Directory blocks
Check vector size**
Op sys tracks
=1 *
=2
=3
=4
256K
51 2K 1 MEG
2 MEG
128
128
128
128
3,7,0
4,15,0
4,15,0
4,15,0
255
255
511
1023
163
127
255
511
OCOh,0
0C0h,0
QFOh,0
OFFh,0
16
32
64
128
=5*
=6
=7
=8
4MFr;
sm2
16ME)G
32ME3G
128
128
128
128
4,15,0
5,31,1
6,63,3
7,127,'
2047
2047
2047
2047
1023
1023
1023
1023
OFFFEh
OEBh,0
QFOh,0
0C0h,0
256
256
256
256
* The numbers one through eight represent partition size,
** HiOos does not allocate check vectors, hence size = 0.
Sectors per track
Blockshift, mask
Block count - 1
Directory count -
Directory blocks
Check vector size
Op sys tracks
8" Floppy
8" Floppy
Single
Double
5" Double sided
Density
Density
Mini Floppy
26
52
32
3,7,0
4,15,0
5,31,0
242
242
156
63
127
127
OCOh,0
OCOh,0
080h,0
16
32
32
2
2
3
Release: 06/21/84
36
Section 4
HiNet Commands
MSDOS Disk Parameters
=1*
=2
=2
=4
256K
51 2K
IMEG
2MEG
128
128
128
128
8
8
8
16
1
1
1
1
2
2
2
2
64
128
256
256
2048
4096
8192
16384
1
2
3
3
6
12
12
128
128
128
128
=5*
=6
=7
=8
4MEG
8MEG
16MEG
32MEG
128
128
256
512
16
32
32
32
1
1
1
1
2
2
2
2
256
256
256
256
32768
65535
65535
65535
4
5
6
7
24
24
12
6
128
128
64
32
Bytes per sector
Sectors per cluster
Reserved sectors
No. FATs
Root dir entries
No. sectors
Media byte
Sectors per FAT
Sectors per track
Bytes per sector
Sectors per cluster
Reserved sectors
No. FATs
Root dir entries
No. sectors
Media byte
Sectors per FAT
Sectors per track
* The numbers one through eight represent partition size.
The first sector of each MSEXDS partition contains a copy of
the above disk parameters. Microsoft refers to this data as the
BPB and it is stored in the format described on page 2-14 of the
MSDOS 2.0 Prograirmer ' s Reference Manual.
Release: 06/21/84
37
Section 4 HiNet Commands
Lock Record
The lock record command is used to lock a record on the
HiNet disk.
poll (50h)
master > user
station 0.16ms station
lock (19h,len,lck)
<
0.38ms
response (stat)
>
. 1 6ms
Total time (including turnaround): 1.7ms
The master returns a single status byte:
Oh = accepted
Ih = denied, lockstring already present
81 h = already locked by current user
2h = denied, illegal string length ( >1 3 or <1 )
82h = denied, lock table full
The total number of lock strings available may be determined by
the Netlnfo protocol. This number is typically 64 for a 5"
master and 128 for an 8" master.
Coomand Parameter
Parameters Length
len = length of lock string (1-13) 1 byte
Ick = lock string 13 bytes
stat= lock status 1 byte
Release: 06/21/84 38
Section 4
HiNet Commands
Unlock Record
The unlock command is used to unlock a record which has
previously been locked.
poll (50h)
master >
station . 1 6ms
unlock (lAh,len,lck)
<
0.38ms
response (stat)
>
. 1 6ms
Total time (including turnaround): 1.7ms
A single status byte is returned:
Oh = accepted
Ih = denied, another user has it locked
2h = denied, illegal string length
82h = denied, lock string not locked
user
station
Conmand Parameter
Parameters Length
len = length of lock string (1-13) 1 byte
Ick = lock string 13 bytes
stat= lock status 1 byte
Release: 06/21/84
39
Section 4 HiNet Commands
Clear Lcxzrks
This command clears all locks which were created by this
user. The "clear locks" command is executed whenever a station
does a warm boot. The locks are also cleared if a station logs
out. Exiting an MSDOS program will not clear locks.
poll (50h)
master > user
station 0.16ms station
clear- locks (iBh)
< .
0.16ms
Total time (including turnaround): 0.82ms
Release: 06/21/84 40
Section 4 HiNet Commands
Hog the Network
The "hog" command allows a user station to gain temporary
control of the network. Immediately after receiving a hog-
acknowledgment, the user station may send or receive any number
of messages on the network, as long as the total time is less
than 16ms. This is enough time to send or receive approximately
1000 bytes. When the user station is finished, it should
relinquish the network by sending an acknowledgment to the
master station. If an acknowledgment is not received within
16ms, the master will assume that the station has failed, and
will attenpt to regain control of the network.
This command allows the user to set up his or her own network
protocols. For example, this command can be used in conjunction
with the WHO command to set up a direct interchange of information
between any two stations on the network.
Note that the Hog protocol does not allow a station to talk
to the master except to terminate the Hog sequence.
poll (50h)
master > user
station 0.16ms station
hog (18h)
<
0.16ms
hog-deny (44h) OR
hog-ack (48h)
_>
0.16ms
(up to 16ms of station-station transmissions,)
(such as Poll Prime, are inserted here)
poll-ack (41 h)
<
. 1 6ms
Total time (including turnaround): 2.14ms to 18.14ms
Release: 06/21/84 41
Section 4 HiNet Commands
Poll Prime
Station to station transmission of 129 bytes of data is
performed in the Z80 and 8086/8088 BIOS by means of the Poll
Prime mechanism. Each user station wishing to receive poll prime
data sets up a poll prime data block and informs the BIOS of its
location (see "Poll Primes and BIOS Calls" document for more
information). The BIOS automatically receives data from pseudo-
user 251. Each user station wishing to send poll prime data must
hog the network from the master, disguise itself as pseudo-user
251 while sending the data to another station, and relinquish the
network back to the master with a poll-ack (41 h) .
*** The hog prologue must be performed here ***
PPpoll (55h)
sending > receiving
station 0.16ms station
PPnack (57h) OR
PPack (56h)
(user 251 ) <-
. 1 6ms
PPdata (129 bytes)
2 . 21 ms
PPnack (57h) OR
PPack (56h)
(user 251) <-
0.16ms
*** The hog epilogue must be performed here ***
Total time (including turnaround): 4.19ms
Release: 06/21/84 42
Section 4 HiNet Corrmands
Get HD Status
This HiNet protocol returns the status of a remote multiple
hard disk (HD) subsystem. Capabilities for passing back status
from both the master's remote hard disk and from a local hard
disk have been provided.
See the accompanying "Poll Primes and BIOS Calls" document
for Z80 and 8086/8088 BIOS calls to receive local and network
hard disk status. The same format 136-byte status data is
returned for all calls.
The protocol is as follows:
poll (50h)
master > user
station 0.16ms station
HD status (iDh)
<
0.16ms
136 bytes: 8 bytes + 128 bytes
>
2.32ms
HD status acknowledge (53h)
<
0.16ms
Total time (including turnaround): 4.30ms
Release: 06/21/84 43
Section 4 HiNet Commands
Get Date and Time
The current date and time maintained by the HiNet master may
be obtained for those stations such as the DMS-1280 and 8086/8088
processors which do not have an internal real-time clock, and is
available to all stations.
The protocol is as follows:
poll (50h)
master > user
station 0.16ms station
date/time (lEh)
<
. 1 6ms
7 bytes: ticks, sees, mins, hrs, mon, day, year
>
. 26ms
date/time acknowledge (54h)
<
0.16ms
Total time (including turnaround): 1.58ms
Release: 06/21/84 44
Section 4 HiNet Commands
Instant Logout
The "logout" command allows a user station to sign off from
the master. The master ensures that the station does not request
a user other than itself to logout, then logs out the station,
releases the station's write ownership partitions and locks, and
deletes any incomplete spool jobs.
poll (50h)
master > user
station 0.16ms station
logout (lFh,usr)
<
0.17ms
logout-deny (44h) OR
logout- ack (41 h)
>
0.16ms
Total time (including turnaround): 1.49ms
Release: 06/21/84 45
Section 4 HiNet Commands
Write Mode
Stations must explicitly obtain write permission of an
ownable partition to ensure that only one user on the network is
capable of writing to that partition at any given moment. The
write mode protocol allows a station to determine ownability of a
partition, take and release that partition, check on its status
and force a change if disaster should strike.
All write mode network calls should be invoked by the BIOS
thru the BIOS jump vectors provided, because additional
information regarding which logical drive letter owns a partition
is stored and checked in each station.
poll (50h)
master > user
station 0.16ms station
write mode (20h,wmc,vol,dsk,val,usr)
<
0.24ms
cmddeny (4Fh) OR
mesack (4Dh,s tat, name)
0.30ms
Total time (incl. turnaround, excl. disk write): 1.7ms
CGnniand Parameter
Parameters Length
vmc = write mode specific ccanmand 1 byte
vol = volume number 1 byte
dsk = unit number 1 byte
val = status value or logical user number 1 byte
usr = physical user number 1 byte
Stat = returned write mode status 1 byte
name = user name 8 bytes
The write mode specific conmands are as follows:
1 = grant ownership to user <val>
2 = release ownership frcxn user <val>
3 = force write status to <val>
4 = query return current write status
5 = clear all write ownership for user <val>
Release: 06/21/84 46
Section 4 HiNet Commands
The returned write mode status values are:
FFh = not owned
FEh = unit is read/write
FDh = unit is read-only
FCh = unit is HiDos shared
FBh = multi-grant (owned by you)
FAh = generic error, illegal parameters
XX = unit is owned by user xx
Release: 06/21/84 47
Section 4 HiNet Commands
Network Information
The "netinfo" command allows a user station to receive up to
128 bytes of information about the configuration of the network
from the master. Several fields are meaningful only when doing
netinfo on the master.
poll (50h)
master > user
station 0.16ms station
netinfo (21 h)
<
0.16ms
netinfo table (1 28 bytes)
>
2.19ms
Total time (including turnaround): 3.51ms
Netinfo table format:
1 byte table version number, (currently 1)
1 byte number users max, (20h or 40h)
1 word* address login user table
1 byte currently
1 byte number spool table entries, (currently lOh)
1 word* address spool table
1 byte number of lock strings max, (40h or 80h)
1 word* address lock string table
remaining entries to 128 bytes reserved
* meaningful only when running on the master
Release: 06/21/84 48
Section 5
BIOS Interface
Section 5: The HiNet BIOS Interface
The HiNet BIOS has several routines which can be called from
a user program to send or receive data on the HiNet cable.
Utility programs such as WHO, DIRNET, and ASSIGN use these rou-
tines to interface with HiNet. All four routines are accessible
though jump vectors at fixed offsets from the base of the BIOS.
Offset
Name
Description
6Fh
SENDNET
Transmit a block of data on the network.
Input:
HL = address of data to be transmitted
BC = number of data bytes
E = pre-transmission delay (master only)
A = user number of intended recipient
72h
RBCNET
Receive a block of data frcm the network.
Input:
HL = address where data is to be stored
BC = maximum number of data bytes
DE = timeout delay (master only)
A = user number of recipient
Output :
A = result status
bit7=0 if timeout (block not received)
=1 if block received
bit 6=0 if no CRC error
=1 if CRC error
bit 5=0 if no receiver overrun
Stations Only>
=1 if receiver overrun
bitO=0 ifa valid poll was not received
1 if a valid poll was received
(User station only)
75h NACKPOLL Wait for next poll, then deactivate auto-
matic poll acknowledgments.
Output :
A = result status (same as RBCNET)
(Master
75h
station only)
INTERCEPT Intercept and process a command from a
user station.
Input :
HL = address of HiNet command
Output :
A = if OK, non-zero if error
The Jump vectors are continued on the next page.
Release: 06/21/84
49
Section 5 BIOS Interface
(User station only)
78h ACKPOLL Reactivate automatic poll acknowledgments.
(Master station only)
78h INTERRUPT Process a one-second interrupt.
Offset
Name
Descr]
iption
84h
HDstat
Check
status
of
local
hard
disk.
(User
station c
)nly)
87h
HDstat
Check
status
of
Network volume (s) .
Programs at any station other than the master can communi-
cate on the network by calling the ACKPOLL routine first (to
synchornize in case of a previous error), then the NACKPOLL
routine. NACKPOLL either receives a poll or returns to the user
after a four-second wait. When no poll is received NACKPOLL
returns with an error status in A. The proper way to test status
is to test each bit individually, or AND the returned value with
the significant status bits: OElh, and test for poll received
with no error: 81 h. When a poll _is received, the program can
then call SENDNET and RBCNET to complete the desired transaction.
When the transaction has been completed, the program should call
ACKPOLL. ACKPOLL will force the BIOS to acknowledge polls auto-
matically until NACKPOLL is called again.
When using SENDNET or RBCNET at a user station, one does not
need to specify a pre- transmission or timeout delay. The pre-
transmission delay will always be 500 microseconds while the
timeout delay will always be about four seconds. The pre-
transmission delay should give the intended recipient of a mes-
sage more than enough time to re-program the DMA and SIO chips.
So, for example, even if several interrupts (timer, spooler)
occur while the chips are being reprogrammed, the recipient has
300 microseconds more than the usual 200 to prepare for reception
of data from HiNet. The master can vary the delays to minimize
wasted time.
Release: 06/21/84 50
Section 5
BIO S Interface
The INTERCEPT and INTERRUPT jump vectors are available only
on the master. The master calls the INTERCEPT routine through
the jump vector whenever it receives a command from a user sta-
tion which has a command byte of (the first byte of a command
is called the "command byte"). Commands can be up to 15 charac-
ters long, including the command byte. By replacing the
INTERCEPT jump vector, users can supply their own command proces-
sing routine to communicate directly with the master station
programs, other user stations, or anything they choose to do
through their applications program.
The code to replace the jump vector might look something
like this:
di ; don't allow interrupts v^ile changing
Ihld 1 ; get address of base of BIOS + 3
Ixi D,73h ; offset of INTERCEPT address
dad D ; compute address of INTERCEPT vector
Ixi D, INTERCEPT
mov M,E ; replace low byte of INTERCEPT vector
inx H
mov M,D ; replace high byte of INTERCEPT vector
ei ; interrupts OK now
; Here is the new INTERCEPT routine
INTERCEPT:
mov
a,m
ora
a
mvi
a,Offh
rnz
inx
H
mov
A,M
•
sub
A
ret
; ensure first command byte was zero
; ret a = Offh if command wasn't zero
; skip first command byte
; look at second byte of command
return status of OK (non-zero if error)
Note that the INTERCEPT routine is invoked also by the
master for illegal commands; so, the user should return a = ffh
for any non-zero command.
The following example is an assembly language program which
shows how a user station can access the who table and the spool
file table. Both are maintained by the master. These tables can
be used to determine who is logged into the network, and what
each current user is doing. Note, the user number is stored in
Release: 06/21/84
51
Section 5 BIOS Interface
the same location at every Z80 station: 47H. This number should
be used to communicate with the master; it should never be
changed .
; Determine size of who table
Beg inWho :
call WaitPoll ; wait for network poll
jrnz BeginWho ; start over again if error
Ixi H,NetInfo
Ixi B,l
sub A
call SENDNET
point to command
one byte command
master is user
send Netlnfo to master
Ixi H,InfoTab ; place to store Netlnfo table
Ixi B,128 ; length of table
call Receive ; Receive from network
jrnz BeginWho ; Try again if failed
Now set up to get Who table
call WaitPoll
jrnz BeginWho ; start over again if error
; Send who command to master
Ixi H,WHOCMD ; point to command
Ixi B,l ; this is a 1 byte command
sub A ; the master is always user
call SENOSIET ; send "who" command to master
; Get who table from master
Ihld InfoTab+1 ; number of users x 1 6 + 1
call Mull 6
inx B
Ixi H,WHOTAB ; address of table
call Receive ; Receive data from network
jrnz GetWho ; Try again if failed
t
; Get spool file table
Ihld InfoTab+5 ; number of spool entries x 16
call Mull 6
Ixi H,SPLTAB ; address of table
call Receive ; Receive data fron network
jrnz GetWho ; Try again if failed
?
; Who table received successfully
CALL ACKPOLL ; re-activate automatic poll ack
Release: 06/21/84 52
Section 5
BIOS Interface
NoTimeOut =
crc_ovr
ValidPoll
NetUsr
= 80h
= 60h
= Olh
= 47h
set if message received
set if crc or overrun error
set if valid poll received
location in menory of user number
General Wait to receive poll routine, returns TZ if poll
WaitPoll:
call Z^KPOLL ; resume acking polls
call NACKPOLL ; intercept the next one
ani NoTimeOut+ValidPoll+crc_ovr ; check status
cpi NoTimeOut+ValidPoll
rz ; return if succeeded
push PSW
call ACKPOLL
mvi C,conouts
Ixi D,NOPOLL
call BDOS
pop PSW
ret
save error status for analysis
resume autanatic poll acknowlegenent
poll not received, so ...
print error message
use the BDOS print function
return A reg and FZ
General Receive network routine, returns TZ if receive OK
Receive:
Ida NetUsr ; this station's user number at 47H
call RECNET
ani NoTimeOut+ValidPoll+crc_ovr ; check status
cpi NoTimeOut ; should get non-poll msg rec'd
rz ; return if succeeded
r
push PSW ; save error status for analysis
mvi c,conouts
Ixi d,xmi trailed
call BOOS
pop PSW ; return A reg and FZ
ret
r
; Multiply L register by 16 and move result to BC
Mull 6:
mvi H, ; zero upper byte
dad H ; multiply by adding to self 4x
dad H
dad H
dad H
push H
pop B
ret
; move result to BC
Release: 06/21/84
53
Section 6 SENDNET & RECNET Listings
Section 6: SENDNET And RBCNET Listings
The SENDNET and RECNET subroutines are used to send or
receive data on HiNet. Below is information on the subroutines.
Listings of these programs follow.
1 . The DMA chip is used to transmit or receive two or more data
bytes, but is not used for one data byte. There are two
reasons for this. 1 : The DMA chip cannot be programmed to
process only one byte. 2: Poll acknowledgments use only one
byte because the DMA chip may be busy doing floppy disk
operations when a poll interrupt is received.
2. The routines are written so that they will work properly on
the DMS-3 and the DMS-4. The DMS-4 is approximately 20%
slower than the DMS-3 due to Multibus access conventions.
Some of the HiNet code is timing sensitive, and it is
probably necessary to rewrite it if one wants to run these
programs under any other timing conditions.
3. In SENDNET, an interrupt is generated when the DMA chip has
finished transmitting all data bytes. The SIO chip cannot be
turned off immediately because it still needs to transmit the
two CRC bytes and several closing flags. The SENDNET routine
waits for the CRCs to be transmitted, then it delays for a
few dozen microseconds to force several closing flags. Now
the SIO chip is turned off, stopping the transmission of
flags.
4. In RBCt-JET, an interrupt is generated when the first character
of a message is received. If several data bytes are
expected, the DMA chip, which will handle the reception of
all remaining data bytes, is activated. Either the DMA chip
or the SIO chip can cause an interrupt, at the end of a
message. Both interrupts will vector to the same routine
(RecLast) , which will save the SIO status register and reset
the SIO. If DMA is being used, the actual data length
received by DMA will be saved (at which time the DMA chip is
reset) for error checking later. If auto- poll acknowledgments
are enabled, the poll is immediately acknowledged; otherwise,
the status byte is set so that the RECNET code can know that
a message has been received.
*Multibus is a tradenark of Intel Corporation, Inc.
.page
Release: 06/21/84 54
Section 6 SENDNET & RECNET Listings
.sbttl 'SENDNET'
++++++++++++++++++++++++++++++++++++++++++++++
+ +
+ Network I/O Routines +
+ +
+ last modified> 12june84 jlw +
+ +
++++++++++++++++++++++++++++++++++++++++++++++
/
; Send a message to the master
; Regs in:
A = message to be
sent
; Regs out:
none
; Destroyed:
A, BC, DE,
HL
SENDMSG:
Ixi
H,NETmsg
mov
M,A
Ixi
B,l
sub
A
master
number
=
•
/
fall thru to SENDnet
Transmit a block on the network
Regs in: HL = block address
EC = byte count
E = delay time (master only)
A = user number
Regs out: none
Destroyed: A, EC, DE, HL
SendNet is extremely time-critical, particularly
for DMS-4's, so modify it at your own risk.
SENDNET:
.ifn MASTopt,[
mvi e,$halfms]; delay approx 1/2 mSecs
inr E ; delay so that receiver has
..spin:dcr E ; time to prepare for message
jrnz ..spin
mov D,A ; save user number
shld DMANSadr; store block address
dcx B ; DMA chip wants real size - 1
shod DMANSsize
mov A,C
ora B
di ; no ints while reprogramming
jrz ..notDMA
; send a data block using DMA
Release: 06/21/84 55
Section 6
SENDNET & RECNET Listings
..useCMA:
Ixi
shld
mvi
out
H,DMATdone; set up the DMA vector
DMAvect
A,l ; multiplex SIOIB to DMA
PIOAD
Ixi
Ixi
outir
H,DMANSprog; program the DMA chip
B,DMAN$<8+DMA
Ixi
Ixi
outir
H,SENDprog
B,SEND$<8+SI01BC
mvi
djnz
Ixi
mvi
mov
di
out
mvi
out
mvi
out
ei
..testDONE:
mov
ora
jrz
jmpr
b,9
; force 3 leading flags
h,sDMAflag
m,0 ; reset dma done flag
Sfd
siolbd ; send user number
a,enaDMA
DMA ; enable dma chip
a,rstEOM
siolbc ; reset underrun flag (SIO)
a,m
a
..testDONE
..fin
; DMATdone sets sDMAf lag
; send one data byte so no need to use DMA
..notDMA:
Ixi H,SENDprog
Ixi B,SEND$<8+SI01BC
outir
. .wait:
mvi
djnz
Ibcd
Id ax
mov
mvi
di
outp
mvi
out
in
b,9
; force 3 leading flags
DMANSadr
b
e,a ; get message byte
c, siolbd
D ; send user no
a,rstEOM
siolbc ; reset under run flag
siolbc
Release: 06/21/84
56
Section 6 SENDNET & RECNET Listings
bit TxRdy,a
jrz ..wait
r
outp E ; send message byte
; Handle the end of the transmission
. .fin:
mvi B,20h ; force CRC bytes
djnz . ; and min 3 closing flags
mvi A,rstCHANNEL
out SIOIBC ; reset the SIO chip
ei
ret
.page
.sbttl 'RecTime, RecMsg, RecNet, and NackPoll'
RBCtime sets the timeout interval for RECNET
Only assembled for a HiNet station
Regs in: BC = timeout constant
Regs out: none
Destroyed: none
.ife station, [
RBCtime:
sbcd Rtime
ret
Rtime:
.word $4 sec
] ; end ' station '
Receive a message from the master
Regs in: none
Regs out: A = xmission err status
NETmsg = receive message char
Destroyed: BC, DE, HL
RECMSG:
Ihld USRadr
mov a,m ; get user number
Ixi H, NETmsg
Ixi B,l ; fall thru to RBCnet
Receive a block from the network
Regs in: HL = block .address
BC = maximum byte count
DE = timeout count (master only)
A = user number
Regs out: A = error status
Release: 06/21/84 57
Section 6
SENDNET & RECNET Listings
bit 7 reset = time-out
bit 6 set = CRC error
bit 5 set = receiver overrun
bit set = poll only rcv'd
Destroyed: A, BC, DE, HL
RECNET:
call RBCbegin ; program the SIO chip
Wait for next poll from network and don't ack
ENTRY> none
EXIT> see RBCnet
Destroyed: BC, HL
NACKpoll:
ei ; make sure SIO can interrupt
Ixi H,ACKflag ; point to ack/nack status
mvi M,l ; don't ack polls (RBClast)
Ixi h,RECstat ; point to receiver status
mvi m,60h ; init RBCstat to errors
.ifn MASTopt,[
Ided Rtime
]
if slave, timeout is 4 sec
; Wait for block received from network cable
..wait: mvi
..loop: mov
bit
rnz
djnz
B,8
A,M
4,A
..loop
; inner loop count
; check receiver status
; return if block received
dcx D ; decrement timeout count
mov A,E
ora D
jrnz ..wait ; fall through if timeout
di
call
res
mov
ret
; prevent real SIO int
REClast ; pretend we received a block
pollRCV,m ; no valid poll rcv'd
a,m ; return error status
Program the SIO to acknowledge polls from the master
Regs in: none
Regs out: none
ACKpoll :
xra a
sta ACKflag ; let RecLast acknowledge polls
SIO programming sometimes fails to take so inside
Constat keep a counter and force a call to AckPoll
Release: 06/21/84
58
Section 6 SENDNET & RECNET Listings
if it wraps around - reset the counter upon every
call to AckPoll.
.ife station, [
sta failed2ack
] ; end ' station '
; fall into begAck:
; starts the poll acking sequence
begACK:
Ixi h,NETmsg; put poll here
Ixi b,1 ; receive one byte
Ida NETusr ; addressed to us
; Program the SIO chip to receive a block
RECbegin:
sta RBCadr ; station address
shld DMANRadr; DMA address
dcx B
sbcd DMANRsize; DMA rev len = real len-1
mov A,C
ora B
di ; no ints while prog vectsl
jrz ..prog
r
; Receive more than one byte
Ixi H,RBClast; set up the DMA vector
shld DMAvect
mvi A,l ; multiplex SIO IB to DMA
out PIOAD
Ixi H,DMANRprog; program the DMA chip
Ixi B,DMAN$<8+DMA
outir
. .prog:
Ixi H,RECfirst; set up interrupt vectors
shld SIOl vect+4
Ixi H,REClast
shld SIOlvect+6
Ixi H,RECprog; program the SIO chip
Ixi B,RBC$<8+SI01BC
outir
ei ; make sure interrupts enabled
ret
.page
Using the prom GONIN and CObDUT routines during
the net boot prevents accessing location 47h (NETusr)
USRadr is initialized to NETusr because it
assembles with station code as well.
24MAR82MdG
USRadr:
Release: 06/21/84 59
Section 6
SENDNET & RECNET Listings
.word NETusr
Nbootusr :
.byte
ACKflag:
.byte
sDMAf lag :
.byte
rcvEMAlen :
.word
RBCstat :
.byte
60h
REC.SP:
.word
.ife Station, [
RcvIntErrs:
.byte
failed2ack:
.byte
]
.page
.sbttl
; User number is here
; during net boot
; ACKPOLL/tlACKPOLL flag
= DMA not done
; actual rcv'd block length
; rev status - in it to timeout,
; crc & ovr err - no poll rcv'd
; SP before interrupt
RecLast error count
; counter for forced PollAck
; end ' not master '
'SDLC interrupts'
; SDLC receive interrupt on first char
RECfirst:
push PSW
push h
in SIOlBD ; flush user number
Ihld DMANRsize; if len = 0, don't use DMA
mov a,l
ora h
j r z ..noDMA
r
; use the dma chip to receive a block of data
mvi A,enaDMA; enable DMA
out
DMA
jmpr
..ret
..noDMA:
in
siol bd ;
Ihld
DMANRadr
mov
m,a
dad
h ;
in
siolbd ;
get message byte
get buffer address
put message into buffer
done for timing
flush first crc byte
Release: 06/21/84
60
Section 6
SENDNET & RECNET Listings
,ret:
pop
pop
ei
reti
h
PSW
SDLC receive interrupt on last char
REClast:
sspd RBC.SP
Ixi SP,RECstck
push D
push PSW
push h
push b
Ixi
in
mvi
outp
inp
outp
ani
set
sta
h,l <8+rstCHAN
siolbd
c,siolbc
h
a
1
OeOh
4,a
RBCstat
flush 1st crc byte if dma
2nd if not dma
hinet control port
get ready to read reg 1
read receive status
reset siolb channel
mask relevant bits
no timeout
.ifn Master, [Ixi d,..begRet] ; in case of error
Ihld
mov
ora
jrz
Ixi
Ixi
outir
Ixi
Ixi
inir
mvi
out
DMANRsize; if len = 0, don't use DMA
a,l
h
, ..pastDMA
h,ckDMAlen ; program the DMA chip
b,DMAlen$<8+DMA ; to send actual rcv'd
; data length
h,rcvDMAlen
b,2<8+DMA
A,rstDMA
DMA
; store actual rcv'd
; data length at
; rcvDMAlen
; reset the DMA chip
•ifn Master, [Ixi
.pastDMA:
d,..ret]; in case of error
.ifn MASTopt,[
; treat poll prime data block as special case
; - PPack or PPnack based on RecStat
Release: 06/21/84
61
Section 6 SENDNET & RECNET Listings
Ihld PPAadr
mov A,M ; Get PP status byte
cpi 2
jrz ..PPblk ; Jump if we got data blk
if there was a network xmission error
then don't respond to anything
Ida RecStat
cpi 90h ; = EOM, no crc or ovr errors
jrz ..ckSize
Ixi h,RcvIntErrs
inr m ; err Net is used by NetErr
xchg
pchl
If DMA size > 0, msg was neither a valid
poll nor a valid poll-prime
..ckSize:
Ihld DMANRsize
mov a,l ; Non-zero if data block
ora h
jnz ..ret
; Check if msg. was a poll-prime
Ihld DMANRadr
mov a,m ; Get msg. byte
cpi POLLpr
jrnz ..ckPOLL; jump if not poll-prime
. .isPrime:
; Reset internal time-out bit used by RECNET
Ixi h,RECstat
res 4,m
}
; Check to see if we are accepting poll-primes
Ihld PPAadr
mov A,M ; Get poll-prime status
cpi 1
Ixi b,1 ; len of msg (for SendNet)
mvi a,PPusr ; net adr (for SendNet)
jrz ..PrmAck; jump if acking poll-primes
}
Ixi h,NetMsg
mvi m,PPnack
call SendNet ; PPnack the PPusr
jmpr ..beg RET
r
} Update status byte
. . PrmAck :
mvi M,2 ; Still pointing at PP status
Release: 06/21/84 62
Section 6 SENDNET & RECNET Listings
; Send poll-prime acknowledge
Ixi H,NETmsg
mvi M,PPack ; sending poll-prime ack
call SENDNET
; Call RecBegin to receive a PP data block
Ihld PPAadr
Point to PP data area
PP msg. length
Our user number
Reprogram the SIO
..ckPOLL:
; insure we received a poll
cpi poll ; if poll, should we ack
jrnz ..begRET; ensure SIO is re- programmed
; Check to ack poll
. .qACK:
Ixi h,RBCstat
set pollRCV,m ; flag for poll rcv'd
Ida ACKflag
rrc
jrc ..ret ; Jump if no pollack wanted
f
; Ack the poll
Ixi
D,3
dad
D
Ixi
B,PPmlen
Ida
NETusr
call
RECbegin
3mpr
..ret
.ife
Ida
ora
OptFlopAny, [
flopBUSY; don't answer if
a ; flop is in use
jrnz
]
•
..skipACK
; end ' OptFlopAny '
mvi
call
A,pollack
SENDMSG
..skipACK:
call
jmpr
J^Kpoll
..ret
..PPblk:
; Update PP status
Ihld PPAadr
Ida RBCstat
res 4,a ; Pretend no poll rcvd
ori 3 ; Indicate PP data blk rcvd
sta RECstat
mov M,A ; Store PP status
Release: 06/21/84 63
Section 6
SENDNET & RECNET Listings
check for errrors on data block receive
ani 80h+crc$ovr
xri 80h ; toggle timeout bit
mvi a,PPack ; ack reception
jrz ..blkResp
mvi m,l ; ready to ack PPrimes
mvi a,PPnack; nack reception
..blkResp:
Ixi
mov
Ixi
mvi
call
h,NetMsg
m,a ; send PPack or PPnack
b,l ; with msg len = 1
a,PPusr ; to default PollPrime user
SendNet
; Reprogram SIO for polls
..begRET:
call begACK
. .ret:
]
end ' not MASTopt '
pop
b
pop
h
pop
PSW
pop
D
Ispd
REC.SP
ei
; ints ok now
reti
; DMA transmit
complete interrupt
DMATdone :
push
psw
mvi
A,rstDMA; reset the DMA chip
out
DMA ; and lEO 1 ine
sta
sDMAflag
pop
psw
ret
.page
.sbttl
'pseudo numbers, network mes
; *** Pseudo-user numbers
BOOTusr == 254
LOGusr == 253
MIMICusr == 252
PPusr == 251
; boot pseudo user number
; login pseudo user number
; mimic pseudo user number
; poll-prime pseudo user number
; Network messages and commands
poll == 'P' ; poll from master
POLLpr == 'U' ; poll-prime from user
Release: 06/21/84
64
Section 6
SENDNET & RECNET Listings
pollack
'A'
poll- received ack
PPack
==
V ;
poll-prime- received ack
PPnack
==
'W
poll-prime received nack
datack
==
'D' ;
data received ack
me sack
==
'M' ;
message received ack
GmdDeny
==
4fh
; cotroand deny
nack
==
•N'
negative acknowledge
logack
==
'L' ;
login acknowledge
lognack
==
'N' ;
login conflict
logdeny
==
'D' ;
login denied
lockack
==
; lock acknowledge
lockdeny
==
1
; lock denied
locknack
==
2
} lock error
not locked
==
82h
; unlock not found
lockfull
==
82h
; lock table full
yourlock
==
81 h
; lock already locked by you
hogack
==
'H'
hog acknowledge
hogdeny
==
'D' ;
hog denied
EiSTack
==
•S'
HD status rcve ack
DTMack
==
Iml
date/time rcve ack
loutack
==
'A'
; instant logout ack
loutdeny
==
'D'
; instant logout deny
whoNhT
==
lOh ;
who command
readNET
==
llh
; read command
writeNET
==
12h
; write command
loginNET
==
13h
; login command
startNhiT
==
14h
; start spool command
AckSpStart
==
04h
; Ack spool start req
read! MET
==
15h
; read IK command
stopNET
==
16h
; stop spool command
assnNET
==
17h ;
assign command
hogNET
==
18h
; hog command
lockNET
==
19h ;
lock command
unlockNE
==
lAh
} unlock command
clrlockN
==
iBh j
clearlocks command
spoolNET
==
iCh
'r spool command
hdstNET
==
IDH ;
hard disk volume status cmd
dttmNET
==
lEh
; date/ time status
loutNET
==
IFh
; instant logout
WrtMode
==
20h
; generic Mode request
infoNet
==
21h
; request general net info
.page
•
; Network
protocols:
MASTER
SLAVE
Poll:
poll
— >
< — polla
-k
Write:
poll ~>
Release: 06/21/84
65
Section 6
SENDNET & RECNET Listings
<— writeNET
mesack
— >
<~ DATA
datack
— >
; Read:
poll — >
< — readNET
DATA
— >
< — datack
; Login:
poll
— >
< — loginNET
logack
— >
(logdeny)
; Assign:
poll
— >
< — assignNET
DATA
— >
; Hog:
poll
-->
< — hogNET
hogack
— >
(hogdeny) .
< — pollack
.page
.sbttl
.prntx
; SIO commands
SENDprog :
.byte
.byte
.byte
.byte
.byte
.byte
SEND$
'sio and dma commands for network activity'
'made it to sio and dma commands'
ooonooob
14h, 001 00000b
3,00100000b
llh,l 10001 00b
5,11101011b
10000000b
channel reset
SDLC mode
auto enables
no interrupts
transmitter enable
reset CRC generator
.-SENDprog
RECprog :
.byte
.byte
.byte
.byte
.byte
.byte
.byte
RECadr:.byte
.byte
.byte
.byte
RBCable:.byte
re::$
ooonooob ; channel reset
4,00100000b ; SDLC mode
2,SI01vect&OFFh ; interrupt
vector
OlOOOOOOb
15h, 10000000b
3,11111100b
6
7,01111110b
llh,11101100b
23h
1 1 1 1 1 1 01 b
reset crc check
transmit disable
receiver info
user number
flag byte
interrupt on
first byte
enable receiver
.-RECprog
Release: 06/21/84
66
Section 6
SENDNET & RECNET Listings
; DMA commands
DMANSprog:
.byte
0C3h
; master reset
2D
.byte
0C7h ;
reset port A
2D
•byte
OCBh
; reset port B
2D
.byte
79h
} write
DMANSadri.word
] filled by SENDNET or RBCstart
DMANSsiz:.word
; filled by SENDNET or RECstart
.byte
14h ;
port A inc, memory
IB
.byte
28h ;
port B fixed, I/O
IB
.byte
95h
; byte mode
2B
.byte
SIOIBD
] port B
.byte
12h
} interrupt at end of block
.byte
DMAvectS
DFFh ; interrupt vector
.byte
92h
; stop at end of block
.byte
OCFh ;
load starting address
2C
.byte
writeDMA
; write
.byte
OCFh
; load starting address
lA
.byte
OABh
} enable interrupts
2D
DMAN$ ==
.-DMANSp
rog
DMANRprog :
.byte
0C3h
; master reset
2D
.byte
0C7h ;
reset port A
2D
.byte
OCBh
; reset port B
2D
.byte
7Dh
; read
DMANRadr:.word
; filled by SENDNET or RBCstart
DMANRsiz:.word
; filled by SENDNET or RECstart
.byte
14h J
port A inc, memory
IB
.byte
28h ;
port B fixed, I/O
IB
.byte
95h
; byte mode
2B
.byte
SIOIBD
; port B
.byte
12h
; interrupt at end of block
.byte
DMAvectS
OFFh ; interrupt vector
.byte
92h
; stop at end of block
.byte
OCth ;
load starting address
2C
.byte
readDMA
; read
.byte
OCFh
; load starting address
lA
.byte
OABh
} enable interrupts
2D
ckDMAlen:
.byte
Obbh
} set read status mask
2D
.byte
06h ;
read the byte counter
2D
.byte
0a7h
; initiate read sequence
2D
DMAlen$
==
.-ckDMAlen
Release: 06/21/84
67
