BSD 4_3_Net_2 release
[unix-history] / usr / src / share / doc / smm / 01.setup / tahoe / 1.t
index 322b1b0..bf3db47 100644 (file)
@@ -1,19 +1,35 @@
 .\" Copyright (c) 1988 The Regents of the University of California.
 .\" All rights reserved.
 .\"
 .\" Copyright (c) 1988 The Regents of the University of California.
 .\" All rights reserved.
 .\"
-.\" Redistribution and use in source and binary forms are permitted
-.\" provided that the above copyright notice and this paragraph are
-.\" duplicated in all such forms and that any documentation,
-.\" advertising materials, and other materials related to such
-.\" distribution and use acknowledge that the software was developed
-.\" by the University of California, Berkeley.  The name of the
-.\" University may not be used to endorse or promote products derived
-.\" from this software without specific prior written permission.
-.\" THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR
-.\" IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED
-.\" WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\" are met:
+.\" 1. Redistributions of source code must retain the above copyright
+.\"    notice, this list of conditions and the following disclaimer.
+.\" 2. Redistributions in binary form must reproduce the above copyright
+.\"    notice, this list of conditions and the following disclaimer in the
+.\"    documentation and/or other materials provided with the distribution.
+.\" 3. All advertising materials mentioning features or use of this software
+.\"    must display the following acknowledgement:
+.\"    This product includes software developed by the University of
+.\"    California, Berkeley and its contributors.
+.\" 4. Neither the name of the University nor the names of its contributors
+.\"    may be used to endorse or promote products derived from this software
+.\"    without specific prior written permission.
 .\"
 .\"
-.\"    @(#)1.t 1.4 (Berkeley) 3/7/89
+.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
+.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+.\" SUCH DAMAGE.
+.\"
+.\"    @(#)1.t 1.6 (Berkeley) 5/7/91
 .\"
 .ds lq ``
 .ds rq ''
 .\"
 .ds lq ``
 .ds rq ''
 .NL
 .PP
 This document explains how to install the Berkeley
 .NL
 .PP
 This document explains how to install the Berkeley
-version of UNIX for the Tahoe on your system.  While this is the first
-release from Berkeley for the Tahoe, the version of
+version of \*(Ux for the \*(Th on your system.  While this is the first
+release from Berkeley for the \*(Th, the version of
 .UX
 distributed by Computer Consoles Inc. (CCI) was derived from 4.2BSD.
 Consequently, the filesystem
 format is compatible and it will only be necessary for you to perform
 a full bootstrap procedure if you are installing the release on a new
 machine.
 .UX
 distributed by Computer Consoles Inc. (CCI) was derived from 4.2BSD.
 Consequently, the filesystem
 format is compatible and it will only be necessary for you to perform
 a full bootstrap procedure if you are installing the release on a new
 machine.
-The System V UNIX systems distributed by CCI, Unisys (Sperry) and Harris
+The System V \*(Ux systems distributed by CCI, Unisys (Sperry) and Harris
 are also derived from 4.2BSD, but only the network and filesystem
 remain compatible with \*(4B.
 The object file formats are completely different in the System V
 are also derived from 4.2BSD, but only the network and filesystem
 remain compatible with \*(4B.
 The object file formats are completely different in the System V
@@ -73,12 +89,12 @@ the vendor-derived versions of 4.2BSD and the Berkeley \*(4B release, many
 4.2BSD binary images will not function properly.  For such programs it
 will be both necessary and desirable to recompile this software after
 the conversion.  Consult section 3 for a description of the differences
 4.2BSD binary images will not function properly.  For such programs it
 will be both necessary and desirable to recompile this software after
 the conversion.  Consult section 3 for a description of the differences
-between \*(4B and the previous vendor-supplied systems for the Tahoe.
+between \*(4B and the previous vendor-supplied systems for the \*(Th.
 .NH 2
 Hardware supported
 .PP
 This distribution can be booted on a CCI Power 6/32, Harris HCX-7,
 .NH 2
 Hardware supported
 .PP
 This distribution can be booted on a CCI Power 6/32, Harris HCX-7,
-Unisys (Sperry) 7000/40, or ICL Clan 7 with any disks supported on the VERSAbus
+Unisys (Sperry) 7000/40, or ICL Clan 7 with any disks supported on the \*(Vs
 disk controllers sold by these vendors (SMD/E or VDDC).
 The new CCI SMD/E controller with working scatter-gather I/O
 is supported as well.
 disk controllers sold by these vendors (SMD/E or VDDC).
 The new CCI SMD/E controller with working scatter-gather I/O
 is supported as well.
@@ -92,9 +108,12 @@ FUJITSU 2351 Eagle  CDC 515M
 Maxtor 340M
 .TE
 .DE
 Maxtor 340M
 .TE
 .DE
+The distribution can also be booted on a Harris HCX-9
+using any disk on the HDC disk controller on the \*(Vm,
+although the \*(Vm tapes are not currently supported.
 .PP
 .PP
-The only tape drives supported by this distribution are the
-ones attached to the Ciprico Tapemaster tape controller.
+The only tape drives supported by this distribution are 9-track tape drives
+attached to the Ciprico Tapemaster tape controller.
 .NH 2
 Distribution format
 .PP
 .NH 2
 Distribution format
 .PP
@@ -138,25 +157,30 @@ Hardware terminology
 This section gives a short discussion of hardware terminology
 to help you get your bearings. 
 .PP
 This section gives a short discussion of hardware terminology
 to help you get your bearings. 
 .PP
-The Power 6/32 (and most related machines being shipped) use a VERSAbus
-for all I/O peripherals.  The console processor used for bootstrap and
-diagnostic purposes is also located on the VERSAbus.  The device naming
-conventions described here apply to the console processor; under UNIX
+The Power 6/32 (and most related machines being shipped) use a \*(Vs
+for all I/O peripherals.
+The console processor used for bootstrap and
+diagnostic purposes is also located on the \*(Vs.
+The Harris HCX-9 uses a \*(Vm instead of a \(*Vs; however, the architecture
+is completely analogous, and the following discussion applies with the
+exception of the name of the bus and the name of the disk controller.
+The device naming
+conventions described here apply to the console processor; under \*(Ux
 device naming is considerably simpler.
 .PP
 device naming is considerably simpler.
 .PP
-The VERSAbus is a 32-bit bus that supports devices which
+The \*(Vs is a 32-bit bus that supports devices which
 use 16-bit, 24-bit, or 32-bit addresses (or some combination).
 use 16-bit, 24-bit, or 32-bit addresses (or some combination).
-The type of each address placed on the VERSAbus is indicated
+The type of each address placed on the \*(Vs is indicated
 by an accompanying \fIaddress modifier\fP.  In addition to the
 width of the
 by an accompanying \fIaddress modifier\fP.  In addition to the
 width of the
-address present on the bus, VERSAbus address modifiers
+address present on the bus, \*(Vs address modifiers
 may be used to indicate the privileges of the requesting
 program (.e.g the program is executing in supervisory mode).
 may be used to indicate the privileges of the requesting
 program (.e.g the program is executing in supervisory mode).
-The 6/32's VERSAbus adapter accepts device requests with either
+The 6/32's \*(Vs adapter accepts device requests with either
 16, 24, or 32-bit address modifiers.
 16-bit addresses are used to access control registers
 16, 24, or 32-bit address modifiers.
 16-bit addresses are used to access control registers
-for VERSAbus devices.
-24-bit addresses are used to access up to one megabyte of VERSAbus
+for \*(Vs devices.
+24-bit addresses are used to access up to one megabyte of \*(Vs
 local memory or device shared memory
 as well as the first 15Mb of main memory.
 24-bit addresses are used for DMA by some peripherals,
 local memory or device shared memory
 as well as the first 15Mb of main memory.
 24-bit addresses are used for DMA by some peripherals,
@@ -169,37 +193,37 @@ that of 32-bit devices.
 Devices which do not support full 32-bit
 addressing can be difficult to work with as their limited addressing
 restricts the placement of I/O buffers in main memory.  Unfortunately,
 Devices which do not support full 32-bit
 addressing can be difficult to work with as their limited addressing
 restricts the placement of I/O buffers in main memory.  Unfortunately,
-because the VERSAbus has had limited acceptance, there are
-very few good VERSAbus device controllers available; this has
-resulted in several non-VERSAbus devices being attached to the
-VERSAbus through bus-adapter cards.  Devices of this sort often
+because the \*(Vs has had limited acceptance, there are
+very few good \*(Vs device controllers available; this has
+resulted in several non-\*(Vs devices being attached to the
+\*(Vs through bus-adapter cards.  Devices of this sort often
 support only 20-bit or 24-bit addressing.
 .PP
 support only 20-bit or 24-bit addressing.
 .PP
-From the Tahoe side of the VERSAbus adaptor,
+From the \*(Th side of the \*(Vs adaptor,
 the three address spaces are mapped so as to avoid
 overlaps.  Physical addresses in the range 0xffff0000 to 0xfffffff are
 the three address spaces are mapped so as to avoid
 overlaps.  Physical addresses in the range 0xffff0000 to 0xfffffff are
-used to access VERSAbus devices which use 16-bit addresses.  References
-to this region of the Tahoe address space result in a VERSAbus
+used to access \*(Vs devices which use 16-bit addresses.  References
+to this region of the \*(Th address space result in a \*(Vs
 transfer with a 16-bit address generated from the lower order 16
 bits of the memory address and a ``short addressing non-privileged I/O
 access'' address modifier (0x10).  Addresses in the range 0xff000000 to
 transfer with a 16-bit address generated from the lower order 16
 bits of the memory address and a ``short addressing non-privileged I/O
 access'' address modifier (0x10).  Addresses in the range 0xff000000 to
-0xffff0000 are used to access 24-bit VERSAbus devices, generating a 24-bit
+0xffff0000 are used to access 24-bit \*(Vs devices, generating a 24-bit
 address and a ``standard addressing non-privileged data access''
 address modifier (0x01).
 Within this range, addresses from 0xfff00000 to 0xffff0000 refer
 address and a ``standard addressing non-privileged data access''
 address modifier (0x01).
 Within this range, addresses from 0xfff00000 to 0xffff0000 refer
-to VERSAbus local memory used by devices (such as the VIOC)
+to \*(Vs local memory used by devices (such as the VIOC)
 for shared communication areas.
 Finally, any other address in the
 the primary I/O adapter space, 0xc0000000 to 0xff000000, generates
 for shared communication areas.
 Finally, any other address in the
 the primary I/O adapter space, 0xc0000000 to 0xff000000, generates
-a 32-bit VERSAbus address with an ``extended addressing non-privileged
+a 32-bit \*(Vs address with an ``extended addressing non-privileged
 data access'' address modifier (0xf1).  Note, however, that 32-bit
 data access'' address modifier (0xf1).  Note, however, that 32-bit
-addresses generated by references to this region result in a VERSAbus
+addresses generated by references to this region result in a \*(Vs
 address with bits 31-30 set to 0.  Thus, for example, a reference to
 address with bits 31-30 set to 0.  Thus, for example, a reference to
-a device located at 0xfe000000 would result in a VERSAbus transfer
+a device located at 0xfe000000 would result in a \*(Vs transfer
 with the address set to 0x3e000000.  A complete list of the characteristics
 of the devices supported in the system may be found in Appendix A.
 .PP
 with the address set to 0x3e000000.  A complete list of the characteristics
 of the devices supported in the system may be found in Appendix A.
 .PP
-The console processor has a set of names for devices:
+The console processor on most \*(Vs machines has a set of names for devices:
 .DS
 .TS
 l l.
 .DS
 .TS
 l l.
@@ -227,7 +251,7 @@ the device; it is computed as
 .DS
 8 * \fIcontroller\fP + \fIdevice\fP
 .DE
 .DS
 8 * \fIcontroller\fP + \fIdevice\fP
 .DE
-Thus, controller 0 (by convention the controller located at VERSAbus
+Thus, controller 0 (by convention the controller located at \*(Vs
 address 0xfff2400), drive 0 would have a \fIy\fP value of 0
 while controller 1 (address of 0xfff2800) drive 0 would have a \fIy\fP
 value of 4*.
 address 0xfff2400), drive 0 would have a \fIy\fP value of 0
 while controller 1 (address of 0xfff2800) drive 0 would have a \fIy\fP
 value of 4*.
@@ -240,6 +264,26 @@ The \fIz\fP value is interpreted differently for tapes and disks;
 for disks it is a disk block, and for tapes it is a file number
 on the tape.
 .PP
 for disks it is a disk block, and for tapes it is a file number
 on the tape.
 .PP
+The HCX-9 uses different controllers and terminology:
+.DS
+.TS
+l l.
+disks on HDC controller        dsk
+Xylogics tapes ???
+.TE
+.DE
+Devices are fully specified to the console processor with:
+.DS
+xxx(x,y,z)
+.DE
+where \fIxxx\fP is one of the above names (e.g. \fIdsk\fP).
+The value \fIy\fP specifies the device unit number.
+Thus, controller 0 (by convention the controller located at \*(Vs
+address 0xfff2400), drive 0 would have a \fIy\fP value of 0
+while controller 1 (address of 0xfff2800) drive 0 would have a \fIy\fP
+value of 4*.
+The \fIz\fP value is interpreted as on the other systems.
+.PP
 The console processor has the notion of a \fIdefault device\fP
 to use with file related commands.  The default device is specified
 according to the form shown above.  Further, the console processor,
 The console processor has the notion of a \fIdefault device\fP
 to use with file related commands.  The default device is specified
 according to the form shown above.  Further, the console processor,
@@ -251,19 +295,20 @@ console processor this can lead to serious confusion.  We will
 return to this problem later in section 4; for now you should
 simply be aware of the difference in naming conventions.
 .NH 2
 return to this problem later in section 4; for now you should
 simply be aware of the difference in naming conventions.
 .NH 2
-UNIX device naming
+\*(Ux device naming
 .PP
 .PP
-UNIX has a set of names for devices which are different
+\*(Ux has a set of names for devices which are different
 from the CCI names for the devices, viz.:
 .DS
 .TS
 l l.
 from the CCI names for the devices, viz.:
 .DS
 .TS
 l l.
-VERSAbus disk drives   dk
+\*(Vs (SMD/E, VDDC) disk drives        dk
+\*(Vm (HDC) disk drives        hd
 Cipher tape drives     cy
 .TE
 .DE
 .PP
 Cipher tape drives     cy
 .TE
 .DE
 .PP
-The standalone system, used to bootstrap the full UNIX system,
+The standalone system, used to bootstrap the full \*(Ux system,
 uses device names of the form:
 .DS
 xx(c,d,p)
 uses device names of the form:
 .DS
 xx(c,d,p)
@@ -285,7 +330,7 @@ the ``b'' identifies this as the second partition.
 .PP
 In all simple cases, where only a single controller is present, a drive
 with unit number 0 (determined by its unit plug on the front of the drive)
 .PP
 In all simple cases, where only a single controller is present, a drive
 with unit number 0 (determined by its unit plug on the front of the drive)
-will be called unit 0 in its UNIX file name.  This is not, however, strictly
+will be called unit 0 in its \*(Ux file name.  This is not, however, strictly
 necessary, since the system has a level of indirection in this naming.
 If there are multiple controllers, the disk unit numbers will normally
 be counted sequentially across controllers.  This can be taken
 necessary, since the system has a level of indirection in this naming.
 If there are multiple controllers, the disk unit numbers will normally
 be counted sequentially across controllers.  This can be taken
@@ -293,32 +338,38 @@ advantage of to make the system less dependent on the interconnect
 topology, and to make reconfiguration after hardware failure extremely
 easy.
 .PP
 topology, and to make reconfiguration after hardware failure extremely
 easy.
 .PP
-Each UNIX physical disk is divided into at most 8 logical disk partitions,
+Each \*(Ux physical disk is divided into at most 8 logical disk partitions,
 each of which may occupy any consecutive cylinder range on the physical
 device.  The cylinders occupied by the 8 partitions for each drive type
 are specified initially in the disk description file /etc/disktab
 (c.f. \fIdisktab\fP(5)).  The partition information and description of the
 drive geometry are written in the first sector of each disk with the
 \fIdisklabel\|\fP(8) program.  Each partition may be used for either a
 each of which may occupy any consecutive cylinder range on the physical
 device.  The cylinders occupied by the 8 partitions for each drive type
 are specified initially in the disk description file /etc/disktab
 (c.f. \fIdisktab\fP(5)).  The partition information and description of the
 drive geometry are written in the first sector of each disk with the
 \fIdisklabel\|\fP(8) program.  Each partition may be used for either a
-raw data area such as a paging area or to store a UNIX file system.
+raw data area such as a paging area or to store a \*(Ux file system.
 It is conventional for the first partition on a disk to be used
 It is conventional for the first partition on a disk to be used
-to store a root file system, from which UNIX may be bootstrapped.
+to store a root file system, from which \*(Ux may be bootstrapped.
 The second partition is traditionally used as a paging area, and the
 rest of the disk is divided into spaces for additional ``mounted
 file systems'' by use of one or more additional partitions.
 .PP
 Returning to the discussion of the standalone system, we recall
 that tapes also took three integer parameters.  In the normal case
 The second partition is traditionally used as a paging area, and the
 rest of the disk is divided into spaces for additional ``mounted
 file systems'' by use of one or more additional partitions.
 .PP
 Returning to the discussion of the standalone system, we recall
 that tapes also took three integer parameters.  In the normal case
-where the Cipher tape drive is unit 0 on the first controller, the
-files on the tape have names ``cy(0, 0)'', ``cy(0, 1)'', etc.
-Here ``file'' means a tape file containing a single data stream.
-The distribution tape(s) have data structures in the tape
-files and though the tape(s) contain only 9 tape files, they contain
-several thousand UNIX files.
+where the Cipher tape drive is unit 0 on the first controller
+(the only unit supported by the standalone utilities), the
+files on the tape have names ``cy(0,0,0)'' (or just ``cy(0,0)'',
+``cy(0,1)'', etc.
+Here ``file'' means a tape file containing a single data stream
+terminated by a tape mark.*
+.FS
+* Note that while a tape file consists of a single data stream,
+the distribution tape(s) have data structures in these files.
+Although the tape(s) contain only a few tape files, they comprise
+several thousand \*(Ux files.
+.FE
 .NH 2
 .NH 2
-UNIX devices: block and raw
+\*(Ux devices: block and raw
 .PP
 .PP
-UNIX makes a distinction between ``block'' and ``raw'' (character)
+\*(Ux makes a distinction between ``block'' and ``raw'' (character)
 devices.  Each disk has a block device interface where
 the system makes the device byte addressable and you can write
 a single byte in the middle of the disk.  The system will read
 devices.  Each disk has a block device interface where
 the system makes the device byte addressable and you can write
 a single byte in the middle of the disk.  The system will read