This chapter provides brief descriptions of features that were new to the DIGITAL UNIX system in this release or had changed significantly from previous releases.
The Common Desktop Environment (CDE) is the new default graphical user interface for DIGITAL UNIX. The CDE environment is designed to provide common services across all UNIX platforms, including a consistent user interface for end users and a consistent development environment for application developers across multiple platforms.
CDE on DIGITAL UNIX is based on the X Window System Release 6 (X11R6) and CDE/Motif 1.0 (OSF/Motif 1.2.4), and supplies the following desktop services and applications:
| Window Management | Workspace Management | SessionManagement |
| Calendar | Calculator | MIME-capable Mail |
| Text Editor | Icon Editor | Terminal Emulator |
| Application Integrator | Print Queue Manager | Windowing ksh |
| Keyboard |
CDE is provided in seven software subsets that require a total of 57.81 MB of free disk space for installation. See the Installation Guide for information on the subset names, contents, and sizes.
The CDE kit contains the following migration tools:
mailcv
mail conversion
This utility converts your
dxmail
folders to the conventional mail format
used by CDE
dtmail.
If you plan to use the
mailcv
utility to convert your existing mail folders, back up the folders
before converting them. Do not use the
-d
option with this version of the
mailcv
utility.
dxcaltodtcm
calendar conversion
This utility converts a DECwindows Calendar,
dxcalendar,
database for use with CDE Calendar,
dtcm.
A brief multimedia tutorial of CDE is located on the DIGITAL UNIX V4.0 Associated Products Volume 1 CD-ROM. Once installed, the video tour can be accessed via the application manager in the "Information" folder. The user simply needs to double click the "CDE Video Tour" icon.
The CDE session manager supports X11R6 screen saver extensions and you can now select animated screen savers instead of a blank screen. This release also enables the automatic locking of screens after a specified idle time. Both features can be modified or disabled from the CDE Style Manager menu. Click on the Screen icon, and select the options you want.
This release provides a new Curses implementation that incorporates the following sets of programming interfaces:
This release of DIGITAL UNIX supports Release 6 of the X Window System, Version 11 (X11R6) patchlevel 12. Prior versions of the operating system supported Release 5 (X11R5) patchlevel 26.
The DIGITAL UNIX port of X11R6 supports all the features and functionality of previous releases of DIGITAL UNIX. It also supports all X Consortium standard features of X11R6.
Included in new features are the following protocol extensions:
Gives clients the ability to use requests that are arbitrarily large, rather than being limited to the size restriction of the core protocol. This can result in a significant performance improvement for applications that use large requests.
Enables double buffering, using the new X Consortium standard.
Complete implementation of full XIE 5.0 protocol with a few exceptions.
See the following section.
The XKB (X Keyboard) server extension is new for X11R6 and for DIGITAL UNIX Version . XKB enhances control and customization of the keyboard under the X Window System by providing the following:
In addition, the X11R5 AccessX server extension for users with physical impairments has been incorporated into the XKB server extension. X11R5 applied to versions of DIGITAL UNIX that preceded this release. These accessibility features include StickyKeys, SlowKeys, BounceKeys, MouseKeys, and ToggleKeys, and control over the autorepeat delay and rate.
Several applications that make use of XKB features are also new with
DIGITAL UNIX version 4.0. These applications include
Xdec,
xkbcomp,
xkbprint,
xkbdfltmap,
dxkbledpanel,
dxkeyboard,
and
accessx.
Refer to the reference pages for more information.
Note that the final revision of the X Keyboard Extension, XKB Version 1.0, will be different from XKB Version 0.65, which is shipping with this release. Avoid creating code that directly references the XKB API and data structures. Any X clients created with direct references must be recompiled and relinked when XKB Version 1.0 is shipped in a future release. You may also have to modify your source code.
DXImageview version 1.0 is an OSF/Motif application for viewing a variety of image
file types such as GIF and JPEG. DXImageview supports the Common
Desktop Environment (CDE) recommended integration for applications.
DXImageview is provided in the subset
OSFCDEAPPS400
and is installed as
/usr/bin/X11/dximageview.
The
-help
option displays information on using the application.
The following new or changed commands and utilities are available in this release.
Mtools
software is included in the
OSFDOCTOOLS400
subset. In prior releases, the software was installed by an optional worldwide
support subset.
The
sendmail
utility now allows the user to configure the fuzzy logic for
mail delivery. Previously, if the recipient's address did not
precisely match any of the user names on the host, a best-match algorithm was
applied against the GECOS field in the
passwd
file. If a unique best-match was found, the mail was delivered to this user.
This behavior is now run-time configurable using the
-oG
option on the command-line. See
sendmail.cf(4)
for more information.
The field width for the
Iused
and
Ifree
fields in the output
of the
df
command has been increased to accommodate 12 digits when using
the
-i
switch.
This modification was made to support very large file systems
where the number of inodes could exceed the field width
that was previously set aside for these fields.
To economize on disk space, reference pages are
now shipped in compressed format.
Compressed files were created with the
/usr/bin/gzip
utility. The
man
and
xman
utilities automatically uncompress the reference pages.
If required, you can uncompress reference page files manually
with the
/usr/bin/gunzip
utility.
The
catman
command has also been enhanced to work with compressed
catman
files. All three commands,
man,
xman
and
catman,
still provide support for uncompressed manpages.
The CDE online help viewer also automatically uncompresses
reference pages when they are accessed via a hyperlink in a
help volume.
For more information, refer to the
man(5)
and the
catman(8)
reference pages.
Terminal support has been enhanced to support non-DIGITAL terminals.
Entries have been added to the
terminfo
databases and the
termcap
file to enable this support.
New tools have also been added to assist users in modifying or porting
other
termcap
and
terminfo
entries to DIGITAL UNIX. These include the following:
captoinfo
- Converts
termcap
files to
terminfo
entries.
infocmp
- Uncompiles and, if required, compares
terminfo
entries.
The
tput
and
tic
utilities have also been enhanced.
GNU Emacs has been updated to
Version 19.28. This version is not upwardly compatible with
GNU Emacs Version 18.5, the previous version shipped with DIGITAL UNIX.
Refer to the appropriate GNU emacs documentation in
/usr/lib/emacs/etc.
This release contains Version 1.12I of the Netscape Navigator
World Wide Web browsing program. Invoke the Netscape Navigator from a
CDE desktop icon, located in the CDE Application Manager's Desktop_Apps
group. The Netscape Navigator can be invoked
directly from the command line by running
/usr/bin/X11/netscape.
You can access detailed help on the Netscape Navigator through the
help menus.
A DIGITAL UNIX home page can be found in
/usr/doc/netscape
in the the file named
Digital_UNIX.html.
The home page contains links to helpful documentation,
including a local copy of the
Netscape Navigator User's Handbook
(consider adding this link to your list of Netscape bookmarks).
The DIGITAL UNIX Installation Guide contains information on how to set up Netscape. See Chapter 6, which covers postinstallation setup tasks.
The
dxterm
terminal utility has the following new features:
loginShell
resource
useWMHints
resource
See the
dxterm(1X)
reference page for more information.
Performance Manager is a real-time performance monitor that allows users to detect and correct performance problems. Graphs and charts can show hundreds of different system values, including CPU performance, memory usage, disk transfers, file-system capacity, and network efficiency. Thresholds can be set to alert you to correct a problem when it occurs, commands can be run on multiple nodes from the graphical user interface, and archives of data can be kept for high-speed playback or long-term trend analysis.
A new utility,
utilupdate,
updates the
setld,
and optionally
ris
and
dmu
utilities on servers that are running a Version of
DIGITAL UNIX prior to Version 4.0. It is necessary to run this
utility on the server prior to attempting to serve DIGITAL
UNIX Version 4.0 clients.
The
setld,
ris,
and
dmu
utilities have been modified for Version 4.0 to allow them to work with
the new format of the distribution media. For more information on the use
of this utility, refer to the
Sharing Software on a Local Area Network
manual.
This release introduces the ability to create a standalone bootable tape of the operating system. You can boot from the bootable tape as easily as you can boot from CD-ROM or a RIS area, but without the overhead of selecting or installing subsets. When you restore your system from the bootable tape, you must reconfigure your system using the System Management applications. You will need to adjust system parameters, such as the host name or IP address,
The binaries and shell scripts needed to create and restore a
bootable tape are installed with the base operating system. The
files reside in
OSFBINCOM400
and no other subsets are needed.
OSFBINCOM400
is the Kernel Header and Common Files (Kernel Build
Environment) subset. The files used by the bootable tape utility
include:
/usr/lib/sabt/etc/fstab
/usr/lib/sabt/etc/inittab
/usr/lib/sabt/etc/profile
/usr/lib/sabt/sbin/finder
/usr/lib/sabt/sbin/pickapart
/usr/sys/bin/btcreate
/usr/sys/bin/btextract
/usr/sys/bin/fsmrg
/usr/sys/bin/mksastape
/usr/sys/bin/mktape
/usr/sys/bin/pmerge
/usr/sys/bin/sboot
You use the
btcreate
utility to create a standalone bootable tape. To
extract and restore file systems from tape at the single-user level, you
use the
btextract
utility.
For more information, see the
btcreate(8)
and
btextract(8)
reference pages.
Partition overlap checks have been enhanced or added to the following commands:
newfs
|
ufs_fsck
|
mount
|
The checks ensure that partitions will not be overwritten if
they are marked in use in the
fstype
field on the disk label. The overlap checks also ensure that no
overlapping partition is marked in use.
If a partition or an overlapping partition has an in-use
fstype
field in the disk label, the following commands inquire interactively if
a partition can be overwritten or not:
newfs
|
mkfdmn
|
addvol
|
swapon
|
voldisk
|
voldisksetup
|
Refer to the reference pages for more information.
Partition overlap checks have been generalized by creating two library
functions:
check_usage
and
set_usage.
Two new
fstype
values have been added:
FS_RAW
and
FS_DB.
For example, the library function
set_usage
could be used by database applications to set the
fstype
field of a disk partition that is in use by the database.
Similarly,
check_usage
can be used to determine the usage of a disk partition or any
overlapping partition.
The
scsimgr
utility creates device special files
for newly attached disk and tape devices. This utility is
automatically invoked at system boot time. You
can execute the command to add device special
files for all disk and tape devices attached to a specified
SCSI bus at any time.
See the
scsimgr(8)
reference page for further details.
This release includes a new suite of graphical single system management applications called SysMan. These applications support installation, configuration, daily administration, and monitoring and tuning. Some of the SysMan applications require you to have superuser privileges or offer a restricted interface to non-privileged users.
At installation time, you can use the graphical Installation Setup application to select software and configure disks.
The graphical applications are integrated into the Common Desktop Environment (CDE). The SysMan applications are available from the CDE front panel by clicking on the Application Manager icon and double clicking on the System_Admin group icon.
There are five groups of system management utilities:
The configuration applications modify system files that control various aspects of network connectivity and system configuration.
The daily administration applications manage users, file systems, remote hosts, licenses, and backup; one application assists you in shutting down the system.
The monitoring and tuning applications monitor processes and tune kernel parameters.
The storage management applications manage disk organization and performance.
The tools provide quick status reports.
The following sections list the system management utilities.
This release provides full installations for new systems, and update installations to update systems from DIGITAL UNIX Version 3.2C, Version 3.2D-1, or Version 3.2D-2 to DIGITAL UNIX 4.0.
Two new user interfaces are available for the DIGITAL UNIX full installation process:
The type of interface presented during the full installation is determined automatically based on your hardware. Systems with graphical consoles and 32MB or greater main memory will provide a graphical interface to the installation. Systems with consoles that do not have graphics capabilities or have insufficient memory to support graphical installations provide a text-based interface.
As in previous releases, there are two types of media to install the DIGITAL UNIX operating system onto your system:
However, for Version 4.0,
the CD-ROM contains file systems that are laid out just as
the software would
be installed on the system and contains directly accessible
root,
/usr,
and
/var
areas. This format makes every operating system
command and utility available to the installation process.
This means that UNIX commands required for
recovery procedures such as restoring corrupt file
systems are readily available even if your operating system
is not yet fully functional.
The new format also eliminates the need for the
environment previously known as the Standalone Environment, which
was a primitive, limited operating system environment.
In DIGITAL UNIX Version 4.0, you can perform a default or custom full installation. In previous releases, these installation types were called basic and advanced, respectively.
The installation procedure still provides access to a UNIX shell that lets you recover from serious system problems such as root file system corruption or to perform general file system or disk maintenance tasks before or during the installation. In previous releases, this option was called the System Management option. In Version 4.0, it is known as the UNIX Shell.
The following enhancements have been made to the installation procedure:
root,
/usr,
and
/var
areas is displayed as you select each optional software subset.
The ability to perform a cloned installation is available in this release. A cloned installation enables you to duplicate the file system layout, file system type, and software subset selections from a similar system that has already been installed with DIGITAL UNIX Version 4.0. Installation cloning can only be performed using RIS. If your system is registered to a RIS environment and has a configuration description file (CDF), the installation procedure retrieves the CDF and uses the system configuration information stored in the CDF to configure and install your system.
After a full installation, the SysMan Configuration Checklist will be displayed and can be used to configure your system. The order in which the applications are presented is a logical order in which to configure your system. You can skip any applications that are not needed for the current configuration.
The following graphical applications are also available in the Configuration group under System_Admin in the Application Manager group once the installation is complete:
In addition, the CDE Application Manager has a launchpoint for NIS Setup, which is a command-line interface for setting up Network Information Services (NIS), such as centralized password and user group files.
Each of the graphical configuration applications has a corresponding setup script that can be used through a character-based interface.
The bind configuration utility,
bindconfig,
provides streamlined management of BIND and
DNS (Domain Name Services) services on a host. Bind configuration supports the creation and
management of name service clients and servers and provides the complete range
of configurations permitted by BIND.
Use this application to:
/etc/svc.conf).
named
daemon.
The disk configuration utility,
diskconfig,
manages certain disk operations for the base system.
Use this application to:
The mail configuration utility,
mailconfig,
configures the
sendmail
utility.
Use this application to:
The network configuration utility,
netconfig,
manages network hardware and related
services on a host. This application supports the configuration and
management of networking adapters on a host, and provides maintenance of
network services and related data files.
Use this application to:
gated
routed
rwhod
/etc/routes
/etc/gateways
/etc/hosts
/etc/hosts.equiv
/etc/networks
The NFS configuration utility
nfsconfig,
manages the Network File System (NFS) and
related processes on a host.
Use this application to:
The print configuration utility,
printconfig,
manages the
printcap
file, print daemon specific directories, and files.
Use this application to:
Internationalization configuration is available if the optional worldwide support subsets are installed. See Section 1.13 for more information.
There are eight applications for the daily administration of your system:
The account manager utility,
dxaccounts,
is used to manage user accounts. It operates on both
base security level systems and enhanced security (C2) level systems.
You must have root privileges to modify account databases.
Use NIS to centrally manage user accounts in a network
environment. NIS allows participating systems to share a common set of
passwd
and
group
files.
The account manager allows you manage both the local and NIS account databases. NIS uses a client-server model. To make changes to the NIS databases, you must run the account manager on the machine designated as the NIS server.
The account manager offers views of each of the system databases. By
default, the account manager starts by displaying the Local User view
showing the contents of the
/etc/passwd
file. The supported views
are:
/etc/passwd)
/etc/group)
/var/yp/src/passwd)
/var/yp/src/groups)
Please refer to the account manager online help for more information.
The archiver utility,
dxarchiver,
provides graphical access to the
tar,
cpio,
pax,
and
compress
commands.
Use this application to:
Archiver functions are classified as follows:
The input and output areas provide icon containers and user text areas. You can drag icons that represent files and directories from the CDE File Manager, or icons representing tape and floppy drives from the system information application. Any icon that you can drag can also be dropped into the input icon container. Type the name of a file, directory, or device into the user input text area.
The Archiver requires you to choose whether to use absolute or relative pathnames. The use of absolute pathnames (those written out in full with an initial slash) and relative pathnames (those written without an initial slash) allows the archiver to control the placement of retrieved files. Files in an archive have either absolute or relative paths. Files with relative paths are restored to the current default directory. That is, they are restored to the directory that is displayed in the default directory area.
The file sharing utility,
dxfileshare,
allows the importing and
exporting of NFS file systems. You can run it with or without superuser
permissions. If you are not root, access to system files is restricted so that
you can only view, but not modify, locally exported file systems and perform
mounts to user directories.
Use this application to:
/etc/fstab
file as permanent mount entries.
/etc/export
file.
The host manager utility,
dxhosts,
manages X clients. You can use this
application to change X host permissions for hosts and to execute
X applications remotely on a host. Other applications in this tool
suite use it as a host name resource.
Use this application to:
xterm
).
xhost
access control list.
The license manager utility,
dxlicenses,
provides access to DIGITAL's License
Management Facility (LMF).
This application allows you to:
The shutdown manager utility,
dxshutdown,
shuts down and optionally reboots your
system.
The following options are available:
When the main window expands, the following options are also available:
The system information utility
dxsysinfo,
displays and monitors system resources on a DIGITAL UNIX system.
Use this application to:
The Audit Manager,
dxaudit,
allows you to create audit reports. For more information see the
Security
manual and the
dxaudit(8)
reference page.
The applications described in this section help you to monitor and tune your system.
The kernel tuner utility,
dxkerneltuner,
manages attributes of loadable kernel
subsystems.
Use the kernel tuner to:
The CDE Application Manager provides launchpoints for Storage Management applications in the System Admin group.
The applications that are available depend on what software you have installed. A launchpoint for the Logical Storage Manager is available with the base operating system.
The CDE Application Manager provides launchpoints for Tools in the System Admin group.
The System Management suite includes four launchpoints in this area. Each runs an output-only command in a window that permits you to repeat the command on demand or at intervals.
The launchpoints and their commands are:
iostat
netstat
vmstat
w
tail -f /var/adm/messages
If you want to run another output-only command, you can run one of the commands in this group, then change the current command in the user interface.
This release complies with many new and changes standards.
Refer to the
standards(5)
reference page for more information.
As of this release, DIGITAL UNIX completes the implementation of the POSIX 1003.1b standard interface as approved by the IEEE standards board in September 1993 (IEEE Std 1003.1b-1993, Realtime Extension). See the Guide to Realtime Programming for more information. The new features are described in sections Section 1.8.11, Section 1.8.12, and Section 1.8.13.
As of this release, the DECthreads library
libpthread.so
implements the POSIX 1003.1c standard interface as approved by the IEEE
standards board in June 1995 (IEEE Std 1003.1c-1995, POSIX System Application
Program Interface). The new POSIX (pthread) interface supported with
DECthreads is the most portable, efficient, and powerful programming interface
for a multithreaded environment. These interfaces are defined by
pthread.h.
See the
Guide to DECthreads
for more information.
In this release the DEC C compiler and
libc
implement the ISO C 94
standard. This behaviour is obtained by calling the DEC C compiler
with the
-isoc94
flag.
This feature is not available with the
-oldc
compiler option, which invokes the previous compiler.
This release includes the following enhancements to the development environment.
Tcl/Tk is now available as part of the base operating system. Tcl/Tk is a public domain unencumbered scripting language and graphical tool kit. In addition to Tcl/Tk, a popular extension package, TclX is also included. TclX provides many UNIX extensions to the Tcl command language. Tcl version 7.4, Tk version 4.0, and TclX version 7.4 are included in this release. See the Installation Guide for information on how to identify and install the appropriate software subsets.
The available programs are:
/usr/bin/tcl
- A tcl shell with TclX extensions.
/usr/bin/tclsh
- A hard link to
/usr/bin/tcl.
/usr/bin/wishx
- A Tcl/Tk/tclX shell.
/usr/bin/wish
- A hard link to
/usr/bin/wishx.
/usr/bin/tclhelp
- A graphical help browser for tcl help.
The following changes have beem implemented for DEC C++.
Refer to the DEC C++ Class Library Reference Manual for details on the threadsafe support, including a new Mutex Package.
The division routines within the Complex Library now catch divide-by-zero errors instead of signaling them.
For iostream assignment operators,
there is no longer a memory leak when you use the
*_withassign
assignment operators to initialize an object for which you have called
xalloc().
Previously, the memory allocated for the object by
xalloc()
was lost.
The String extraction operator now takes care of dynamically allocating the String to accommodate the input.
ios::ate
mode
When you open a file specifying
ios::ate
but not
ios::app
to the
filebuf
open()
function, the file is no longer opened in
O_APPEND
mode. This incorrect behavior caused all data to be written to the
end of the file, regardless of the current file position.
Various problems with exception handling have been fixed. Also, support for exception handling in DEC C++ version 5.3 has been added.
exp()
returns zero for underflow errors
When the Complex Library
exp()
function detects an underflow error, the
resulting value is now (0,0) instead of (+/- max-float, +/- max-float).
clog()
and C++ Class Library iostream
clog
A single application is restricted from using both the math library function
clog()
and the iostream package's
clog
object. This restriction is due to
the fact that
libm
and
libcxx
each contain a definition for the global symbol
clog
and those definitions are incompatible.
Furthermore, applications which reference one of the
clog
symbols cannot include both
-lcxx
and
-lm
on their
ld
command line. An error will be generated by
ld
because
clog
is multiply defined.
catch(...)clause
fstream
close()
clears the error state
The
fstream,
ifstream,
and
ofstream
close()
member functions now clear the
stream's error state when the close succeeds.
Call the
clear()
member function after the call to
close().
The previous version of the
pixie
utility has been relocated to the file
/usr/opt/obsolete/usr/bin/pixie
and is provided as part of the
OSFOBSOLETE
subset. The previous version of
pixie
will be retired in a future release and is replaced
with the Atom-based
pixie
utility. A driver for the Atom-based
pixie
is located in
/usr/bin/pixie,
for backward compatibility with the previous version.
The driver accepts the same switches recognized by the
previous version.
Refer to the
pixie(1)
and
atom(1)
reference pages for more information on
pixie
and general information on the Atom tools.
The Software Development Environment (SDE) has been repackaged to ease
installation, simplify licensing, and create a product identity. The current
SDE components have been repackaged into a single, new
OSFSDE
subset and all
of the pieces outside the SDE have been moved into logical subsets, including:
OSFINCLUDE
for all include files
OSFLIBA
for all static libraries
OSFPGMR
for commands outside the scope of the SDE
Because the compiler is
needed at installation time, some SDE components have remained in the
mandatory
OSFCMPLRS
subset.
The Ladebug debugger subsets have been
renamed to the
OSF*
subset name prefix and can now be installed during a
custom installation of DIGITAL UNIX.
These changes have been made on the DIGITAL UNIX Operating System
Volume 1 CD-ROM.
The FUSE Porting Assistant has been added to the DIGITAL UNIX
kit on the
DIGITAL UNIX Associated Products Volume 1 CD-ROM. This is a tool
to help port code to
DIGITAL UNIX from a variety of platforms and operating systems.
The
OSFSDECDE
subset was also added to the
DIGITAL UNIX Operating System Volume 1 CD-ROM. It contains the files
necessary to access DECladebug and the Porting Assistant from CDE.
The C compilers,
cc
and
c89
on DIGITAL UNIX have changed to
DIGITAL's DEC C compiler. The previous default C compiler
is still a supported viable option and can be accessed by
specifying
-oldc
on the
cc
and
c89
command lines.
DEC C uses DIGITAL's back-end compiler (GEM) technology, which has been specifically developed and optimized for use with Alpha systems. Both compilers have full binary compatibility with each other.
The execution order for
init
routines in static executable files has
been modified to more closely match the execution order for
init
routines in dynamic executable files. The init routines
loaded from an archive library will be executed prior to any
init routines loaded from objects and archives occurring
earlier on the linker command line. Prior to this change,
init
routines were executed in the order they were encountered in
processing the
link
command from left to right. As a result,
init order for static executable files was much different than the
init order for equivalent shared executable files.
For existing applications that rely on the static
inito
rder
used in prior releases of DIGITAL UNIX, the new linker option:
-old_init_order
can be used to restore the strict left-to-right
execution order for static executable files.
The
prof
command's pc-sampling mode now supports profiling the shared
libraries used by a program. Linking a call-shared program with the
cc
command's
-p
switch causes the resulting program to profile
both the call-shared executable file and all the shared libraries.
The following command displays a combined profile:
#
prof -all
New
-all,
-incobj,
-excobj,
and
-stride
switches for the PROFFLAGS environment variable
enable you to request per-procedure profiling
of the shared libraries or to select
particular libraries to profile.
Related enhancements are:
monitor(),
monstartup(),
and
profil()
cc
-p
and
cc
-pg
profiling (
gprof),
except for calls to the traditional
monitor()
API.
monitor_signal()
with threads.
prof
and
gprof
checking.
See the
prof(1)
and
monitor(3)
reference pages for further information.
The following
atom
and
prof
commands now profile the shared libraries used by a program:
#
atom -tool pixie -all
and
#
prof -pixie -all
Also, the
threads
environment
for
theatommakes
pixie
tool thread-safe, though per-thread counts are not recorded.
See the
atom
(1),
prof
(1), and
pixie
(5) reference pages for further information.
DIGITAL UNIX Version introduces
the Thread Independent Services (TIS) application
programming interface in the C run-time library
libc.
TIS provides services that assist in the development of
thread-safe libraries.
Thread synchronization may involve significant run-time cost, which is undesirable in the absence of threads. TIS enables thread-safe libraries to be built that are both efficient in the non threaded environment, yet provide the necessary synchronization in the threaded environment.
When DECthreads (pthreads) are not active within the process, TIS executes only the minimum steps necessary. Code running in a non threaded environment does not encounter overhead incurred by the run-time synchronization that is necessary when the same code is run in a threaded environment. When DECthreads are active, the TIS functions provide the necessary thread-safe synchronization.
DIGITAL UNIX Version has an optional high-resolution clock. To enable this option, add the following line to the kernel configuration file and rebuild the kernel:
options MICRO_TIME
The system clock (
CLOCK_REALTIME)
resolution as returned by
clock_getres
will not change. Timer resolution remains the same. However, time as returned
by the
clock_gettime
routine will now be extrapolated between the clock ticks. The granularity of
the time returned will now be in microseconds. The time values returned are
SMP safe, monotonically increasing, and have 1 microsecond as the apparent
resolution.
Realtime signals have been implemented to conform to the POSIX 1003.1b standard. This new feature includes queued signals with optional data delivery, and 16 user-definable realtime signals.
The following functions to support realtime signals were implemented:
sigqueue
sigtimedwait
sigwaitinfo
timer_getoverrun
Synchronized I/O (file synchronization) has been implemented to conform to the POSIX 1003.1b standard. New functions for synchronized I/O under the UFS and ADVFS file systems include:
aio_fsync
Asynchronously writes changes in a file to permanent storage
fdatasync
Writes data changes in a file to permanent storage
The
open
function now takes the following new flags for synchronized I/O:
O_DSYNC
Ensures synchronized I/O data integrity of the file accessed
O_RSYNC
Used for synchronized I/O read operations
For applications conforming to POSIX 1003.1b, the
_POSIX_4SOURCE
macro is
supported for this release, but will be retired with the next
release of DIGITAL UNIX. The macro
_POSIX_4SOURCE
is part of an obsolete draft standard and is supported in this release
for compatibility only. When possible, existing applications which use
_POSIX_4SOURCE
should be modified to
use
_POSIX_C_SOURCE
instead.
The
_POSIX_C_SOURCE
macro is associated with a value, which allows an
application to specify the namespace it requires. However, as a general rule,
avoid explicitly defining standards macros when compiling your applications.
For most applications, the header file
unistd.h
provides the standards definitions needed.
The Digital Porting Assistant is a Motif-based tool to help you port your C, C++, and FORTRAN source code to DIGITAL UNIX from other UNIX and proprietary platforms, including OpenVMS. The Porting Assistant provides the following features:
The Porting Assistant is licensed and provided to you with DIGITAL UNIX Developers' Toolkit but requires separate installation.
To install Version 2.0 of the Porting Assistant, install subsets
PRTBASE200
and
PRTMAN200
(and their dependencies) from the
DIGITAL UNIX Associated Products Volume 1
CD-ROM..
The following networking enhancements have been implemented.
This release includes a new version of the
gated
routing daemon. The
update installation procedure will detect if your system is configured to run
the
gated
routing daemon. If the DIGITAL-supplied
gated
is detected then the
/etc/gated.conf
file is moved to
/etc/ogated.conf.
Otherwise, if a user supplied or customized
gated
is detected, then both the
/etc/gated.conf
and the
/usr/sbin/gated
are saved with the
.PreUPD
suffix.
When the system is installed, the new gated R3.5 is the
default version in
/usr/sbin/gated.
The old
gated
version 1.9 is supplied in
/usr/sbin/ogated.
Also, corresponding older
gated
reference pages are saved with an
o
prefix.
This release contains both a client and a
server Dynamic Host Configuration Protocol (DHCP) daemon.
For DHCP client configuration, use the
netconfig
utility.
For configuration of client parameters on the DHCP server,
use the
/usr/bin/X11/xjoin
utility, which provides a graphical
user interface to the
/etc/bootptab
file.
For more information on DHCP, refer to the
joinc(8)
and
joind(8)
reference pages.
This release supports Point-to-Point Protocol PPP, including
support for BSD-style compression of entire packets.
This is a negotiated option. If a foreign peer cannot handle
this, it should be gracefully rejected via the
Protocol-Reject
of LCP.
When using PPP with modems doing compression, it may be desirable
to force no BSD-style compression. To do this, put
-bsdcomp
in either
/etc/ppp/options,
or on the
pppd
command line.
PPP now has a configurable (at boot time) number of interfaces.
The default is 1. To specify a higher value, add the following line
to the
/etc/sysconfigtab
file and reboot the system:
ppp:
nppp=x
PPP documentation is available in the
pppd(8),
pppstats(8),
and
chat(8),
reference pages and in the
Network Administration
Guide.
This release contains DIGITAL's first implementation of NFS support over the TCP transport. Previously, only the UDP transport was available. UDP is still available and may still be the preferred transport in local area networks. For NFS access over wide area, congested, or lossy networks, TCP may offer better performance.
Use the
nfssetup
script or the
nfsconfig
utility to
enable TCP on the NFS server.
The
nfsiod
daemon has been changed to spawn kernel
threads (instead of forking multiple processes as it
did previously). Each
nfsiod
thread can handle UDP
or TCP mounts, so the command
nfsiod
still accepts one argument.
A new SNMP architecture is present in this release.
The SNMP daemon,
snmpd,
is now an extensible master agent.
End user programmers can develop subagent programs
that communicate with
snmpd
to implement their MIBs on DIGITAL UNIX systems.
The base operating system MIB support is implemented in
a subagent program called
os_mibs,
which is started/stopped automatically with
snmpd.
This release supports the Host Resources MIB (RFC 1514). The MIB support daemon must query the system's devices to retrieve information required for this MIB. This query occurs when the daemon starts, and subsequently whenever a relevant SNMP request arrives.
This device querying is the default behavior, and may be configured
off. See the
snmpd(8)
reference page for more information about configuring
the SNMP agent.
The following new features were implemented:
NEW-ENVIRON
parameter was added to implement RFCs 1571 and 1572 for
telnet
mrouted
was added, including the utilities:
mtrace
map-mbone
mrinfo
The
xntp
utility has been upgraded to Version 3.0, which is a complete implementation
as defined in RFC 1305.
Note that the new
ntpdate
utility requires that the version of the server be specified with the
-o
flag. For example:
%
ntpdate -o 2 server.oxo.fff.com
The default version is
3.
The new
xntpd
daemon
supports authentication. If you choose to enable this
option, note that you only have to provide keys to the servers
and peers that also support authentication.
You should provide names of three servers during setup with
ntpsetup,
choosing the mode
server
and the version option
V2
(In the case where you are not upgrading the servers).
This release provides the following new enhanced security features.
setrlimit
ttys
database is updated on logins.
ttys
has been extended to X displays.
ttys
information are stored in database files for faster access and update
(resulting in faster logins).
edauth
and
convuser
are available.
See the
Security
manual and the
setrlimit(2),
edauth(8),
and
convuser(8)
reference pages.
The following file system enhancements have been implemented in this release.
The following new features were introduced for AdvFS.
There is a new mechanism for limiting the amount of kernel memory that AdvFS uses for its access structures. This may be necessary only for systems with 64 Mb or less memory, and AdvFS as the default file systems. This is applicable to all hardware configurations.
There are two new kernel paramaters relevant to AdvFS that can be
modified using the
sysconfig
or
sysconfigdb
commands. They are
AdvfsAccessMaxPercent
and
AdvfsAccessCleanupPercent.
There is a complete description of these parameters in the
Guide to File System Administration for the POLYCENTER Advanced File System and Utilities for DIGITAL UNIX.
Traditionally, AdvFS directories were never truncated, even though many of the files in the directory had been deleted. This created a problem if the directory file became very big. For example, if several hundred thousand files were added to a directory, then the directory file itself grew very large. Even though most of the files in that directory were subsequently deleted, operations that required scanning the directory remained inefficient because the entire directory file still needed to be read.
AdvFS now truncates directory files when all of the entries at the end of the directory have been deleted. This truncation is done on 8KB byte boundaries, so the size of a directory is always a multiple of 8192.
One ambiguity of directory truncation is that the truncation is done when files are created and not when they are deleted. This is done because of the efficiency of underlying algorithms, and is the same model used by UFS for directory truncation. For example, after most files in a given directory are deleted, the size of the directory file itself will not decrease until a new file is inserted into that directory.
ACLs (Access Control Lists) on files and directories are a new feature
in this release. They are manipulated with the
getacl
and
setacl
commands.
See the
Security
manual and the reference pages for more information.
DIGITAL UNIX Version provides the following new features for the Logical Storage Manager (LSM):
volsave
and
volrestore,
provide an easy way to back up and restore the
LSM configuration database. See the reference pages
for these commands.
dxlsm,
now provides support for disk operations. For example,
adding a disk to LSM.
The functionality and syntax of the LSM commands used for encapsulation, unencapsulation, and mirroring have changed in this release, as follows:
volencap
command now supports the following features
and functions. For details, refer to the
volencap(8)
reference page.
voladvdomencap
command.
c
maps the entire disk, and that
an in-use partition has an
fstype
field other than
UNUSED.
(If a partition's
fstype
field is
UNUSED,
then
volencap
may allocate that partition table
entry for its use.)
volrootmir
command now supports the following features
and functions. For details, refer to the
volrootmir(8)
reference page.
-a
option. This option requires the target disk to be of the
same type as the source disk.
-a
option. This procedure
requires that the target root and swap partitions are
large enough to hold
rootvol
and
swapvol,
but the target and source disks need not be of the same type.
-a
option, the
volunroot
command unencapsulates all
LSM volumes on the system disk, not just
rootvol
and
swapvol.
The requirements for unencapsulation are:
nopriv
disk.
For details, refer to the
volunroot(8)
reference page.
Two new functions,
check_usage
and
set_usage3
are available for use by applications.
These functions check whether a disk partition
is marked for use and set the fstype of the partition
in the disk label. See the reference pages for these
functions for more information.
In DIGITAL UNIX , a new method for configuring and building
kernels is in place. It is designed to meet future needs for modularity,
configurability, maintainability, scalability, and third-party device
installation. Although the internal details of building kernels has
changed significantly, the external interface is almost identical. The
System Administration
guide and
Writing Device Drivers: Tutorial
guide contain more detailed information on this process. For normal
operations, a static
vmunix
is recommended. Bootstrap linking is intended to support
foreign devices, where support for such devices is not delivered in
DIGITAL UNIX.
Previously, the DIGITAL UNIX kernel was comprised of over 1000
individual
.o
files. The configuration procedure,
doconfig,
selected among those files, and compiled a static kernel image that
was then installed and booted. The
.o
files are still available in the base kit and can be loaded separately
if required.
In this release, the
.o
files are grouped into 150-200 modules,
with the suffix
.mod.
Each module represents a functional piece of the kernel, such
as a device driver or platform support module. These modules are the new
form of the original dynamically loadable
modules introduced in DEC OSF/1 Version 3.0.
Modules are selected by
doconfig
instead of the individual files previously used. This simplifies the
Makefiles, saves considerable disk space, and reduces compilation time.
There is a new syntax in the
files
file for making modules:
| Module Definition | Syntax |
MODULE/[STATIC]/modname
|
optional opt1 ....
|
dir/file1.c
|
module modname
|
dir/file2.c
|
module modname
|
dir/file3.c
|
module modname
|
This procedure defines a module:
modname.mod
(note that
.mod
is appended to
modname
automatically). Each module has three files:
file1.c,
file2.c,
and
file3.c.
When each file has been compiled, the module
is linked together, and is ready to build into a kernel.
Optionally,
STATIC
may be specified before the module name.
If it is omitted, the object files are merely linked together into a
module. If
STATIC
is used, the module is assumed to
have a
modname_configure()
routine and
modname_attributes[]
table to make it a configurable and potentially loadable module. Linkages to
these routines are made in the kernel's
static_subsys_list[]
table, and the configure routines will be called at system startup.
See
System Administration
for more information on modules,
making modules loadable, and the use of configure routines.
If an individual file in
BINARY
is rebuilt, the module that contains it must also
be rebuilt to incorporate the change into the kernel. To help with
this, an environment variable
DO_MODULES
is checked. Providing that the value of
DO_MODULES
is the string
AUTO,
a module will be automatically relinked whenever an
individual file is rebuilt.
Kernel developers are encouraged to put their files into modules, and to
consider making them loadable if possible.
To enhance configurability, and help mainstream modules, a new technology called bootstrap linking has been created. This is a way to defer the actual linking of a kernel until boot time. It is provided as an alternative to booting a statically linked kernel.
When you run
doconfig
it builds a static kernel by default. To build a bootlink
kernel, the command option
-b
must be explicitly specified.
A bootstrap linked kernel is actually a set of directives in a
sysconfigtab
file that tell the bootstrap program how to build the
kernel in memory. This file is generated using the make
sctab
rule, which will be invoked by
doconfig.
The updated file is
placed into
/usr/sys/MACHINE/sysconfigtab.
It must be copied into
the root file system, and the two machine-specific modules
MACHINE.mod
and
EXTRAS.mod
copied
into the
/sys/BINARY
directory. The
sysconfigtab
file is then given to the bootstrap program instead of
vmunix
to build and boot the new kernel.
See also the
doconfig
(8) and
sizer
(8)
Reference pages for more information.
There is one major restriction when using a bootstrap linked kernel.
Because there is no
vmunix
executable file, programs that attempt to
search the kernel's symbol table will not find one. This includes
dbx
and
nm,
and most of the compiler tools. Programs that use the
nlist
library call and pass the name of the booted file will
work, even if that file is a
sysconfigtab
file.
To avoid this restriction in DIGITAL UNIX , execute:
#
ld -o <filename> `sizer -m`
This will generate an exact copy of the currently running kernel complete with symbol table in the file <filename>.
When building a bootstrap linkable kernel,
doconfig
first builds a static kernel. If the static kernel fails to build,
do not attempt to bootstrap link the
sysconfigtab
file. Note that this static kernel will not
be identical to the bootstrap-linked kernel as the kernels may be
linked at different addresses. Use the
ld
command if required.
To effect full kernel modularization, several internal changes were made
to the
config
program and the compile environment for kernel files. These
changes are significant for anyone developing kernel drivers or layered
products.
To make all modules the same, a
-dc
linker switch was added to build modules. This has the effect of forcing a
full declaration and storage allocation for all commons (globals). Previously,
if two files both declared a global
int
foo;,
the linker would coalesce them into the same symbol when linking the kernel.
This still happens within a module, but not across modules. Each module would
have its own declaration
foo,
resulting in multiple definitions of the same global. Because
of the multiple definition, kernel linking would fail. To avoid this,
make all but one declaration external, using the
extern.
variable declaration
The
BINARY
file is no longer configurable. All files
that are listed as Binary are compiled, and placed into modules as
defined. The
BINARY
configuration file no longer has any information
about hardware configuration (bus, devices or controllers) or most
options (such as CDFS or NTP_TIME). Neither are any of these header files
built in
BINARY.
BINARY
files that use the
#include
statement to include other files, and the substitution statement
#define,
will no longer compile.
For example:
#include <device.h>
#define NDEVICE 4
This includes files that use the following inclusion statement:
#include <data/NotBinary_data.c>
See the
doconfig(8)
and
sizer(8)
reference pages for more information.
The following new features have been implemented to support internationalization.
The I18N Configuration Tool, available through the CDE Application Manager, is one of the SysMan system administration configuration tools. The I18N Configuration Tool provides a graphical interface that enables you to configure internationalization-specific settings. It also provides a convenient way to see which countries, locales, fonts, and keymaps are currently supported on your system. Use this tool to remove unused fonts and unrequired country support.
This release provides a new set of locales and codeset convertors that support the Unicode and ISO 10646 standards. The codeset convertor modules enable an application to convert between other supported codesets and UCS-4.
DIGITAL UNIX also provides a function called
fold_string_w()
that maps one
Unicode string to another, performing the specified Unicode character
transformations. For more information on the
fold_string_w()
function, see the
fold_string_w(3)
reference page.
For more information on the Unicode support, see the
Unicode(5)
reference page.
Worldwide support subsets no longer install internationalized Mail Handler
(MH) software in the
/usr/I18N/bin/mh
directory. In DIGITAL UNIX Version 4.0,
internationalization features have been merged into the default Mail Handler
(MH) whose files are located in
/usr/bin/mh.
Check the value for the
mhpath
resource used to
find the DECwindows Mail application. If necessary, change this value to be
/usr/bin/mh.
The
mule
editor is a multilingual version of GNU Emacs and supports the
following kinds of characters:
The
IOSWWMULE400
subset installs Version 2.3 of the GNU mule editor and
associated software. Corresponding sources are available in the
IOSWWMULESRC400
subset.
DIGITAL UNIX does not include public domain fonts that you can use with mule.
Refer to the
mule-2.3/README.Mule
file installed by the
IOSWWMULESRC400
subset
to find out how you can obtain public domain fonts.
The DIGITAL UNIX software is enhanced with lisp libraries that support the
dechanzi
codeset for Simplified Chinese and the
dechanyu
codeset and tsangchi
input method for Traditional Chinese. These libraries are included in the
IOSWWMULE400 subset and installed in the
/usr/i18n/mule/lib/mule/site-lisp
directory.
For more information about mule, see the
mule(1)
reference page.
DIGITAL UNIX Version 4.0 includes support for the the
Catalan(5),
Lithuanian(5),
and
Slovene(5)
reference pages for information about associated
codesets, locales, keyboards, and fonts.
The man command can automatically invoke the
iconv
utility to perform codeset conversion of reference page files.
This allows you to install one
set of reference pages to support locales that have the same language and
territory but different codesets, thereby reducing file redundancy on the
system. For more information, refer to the
man(1)
reference page.
PCMCIA (PC Card) support is a new feature in this release and is limited to the following capabilities:
The following restrictions apply in this release.
and the SWAPBOX PREMIUM COMBO ... Model MMCD-FC2:
However, other ISA to PCMCIA bridge adapters using the Intel i82365SL or compatible chip may also work.
Before inserting the PCMCIA adapter board into your system, make sure to read the manual that came with the adapter from the adapter vendor and follow the instructions on how to connect the cables and install the board.
>>>
isacfg -slot 1 -etyp 1 -dev 0 -mk -iobase0 3e0 -irq0 14 -enadev 1 -handle PCIC-PCMCIA
NOTE that if the system is already using slot 1, use other slot number that is not used.
type >>>init to use these changes
>>>isacfg -slot 1
============================================================= handle: PCIC-PCMCIA etyp: 1 slot: 1 dev: 0 enadev: 1 totdev: 1 iobase0: 3e0 membase0: 8000000000000000 iobase1: 8000000000000000 memlen0: 8000000000000000 iobase2: 8000000000000000 membase1: 8000000000000000 iobase3: 8000000000000000 memlen1: 8000000000000000 iobase4: 8000000000000000 membase2: 8000000000000000 iobase5: 8000000000000000 memlen2: 8000000000000000 rombase: 8000000000000000 romlen: 8000000000000000 dmamode0/chan0: 80000000 irq0: 14 dmamode1/chan1: 80000000 irq1: 80000000 dmamode2/chan2: 80000000 irq2: 80000000 dmamode3/chan3: 80000000 irq3: 80000000
============================================================= >>>
Since a PC Card is a dynamic device (i.e. not a static device that is
present all the time in the system hardware), and the serial-line
device driver is a static device driver, when the system is installed
initially, there will not be a corresponding
acex
entry created
automatically by the
doconfig
of the target system. This is
due to the fact that the system doesn't know when it is being
installed that there will be a fax/modem card for PCMCIA since the
card is not in the system yet.
If you want the system to automatically create the 'acex' entry for your PCMCIA fax/modem card, before you start installing the system, make sure that you have the PCMCIA adapter configured in the console and that the PCMCIA fax/modem card is inserted into the slot. If you have a fax/modem card in the slot 0, for instance, when the system is installed and the target kernel is built, the system kernel configuration file built will have the following entry:
controller ace2 at pcmcia0 slot 0 vector aceintr
The installation will also create the device special file for
this fax/modem card in the directory named
/dev.
#
ls -gl tty02
crw-rw-rw- 1 root system 35, 2 Oct 16 13:22 tty02
If you didn't have the PCMCIA fax/modem card inserted in the slot
when the system was installed, then you need to
add the following line to your system kernel configuration file,
/sys/conf/HOSTNAME
where
HOSTNAME
is the name of your system. If you
are just going to use one PCMCIA fax/modem card:
controller ace2 at * slot ? vector aceintr
If you plan to use 2 modem cards simultaneously, add the following 2 lines to your system configuration file:
controller ace2 at * slot ? vector aceintr controller ace3 at * slot ? vector aceintr
Once the system configuration file is modified, use the following command to rebuild the new kernel and reboot the system.
#
doconfig -c
Normally the system installation will create the following two
default
tty0x
device special files in the directory
/dev,
as the following command shows:
#
ls -gl tty0*
This command produces output similar to the following:
crw-rw-rw- 1 root system 35, 0 Oct 16 13:22 tty00 crw-rw-rw- 1 root system 35, 1 Oct 16 13:22 tty01
This is because most systems have 2 imbedded serial lines.
If the system has only one, only tty00 entry will be visible in the
/dev
directory.
Create additional device special files for the PCMCIA modem cards
using
the
MAKEDEV
utility in the
/dev
directory.
For example, in the
/devdirectory:
#
./MAKEDEV ace2
MAKEDEV: special file(s) for ace2: tty02
The generated special file should look like this:
crw-rw-rw- 1 root system 35, 2 Oct 27 14:02 tty02
If you intend to have 2 PCMCIA modem cards working simultaneously, create device special files for the number of cards, in this case, 2. For example, in /dev directory:
#
./MAKEDEV ace2 ace3
MAKEDEV: special file(s) for ace2: tty02 MAKEDEV: special file(s) for ace3: tty03
The generated special file should look like this:
crw-rw-rw- 1 root system 35, 2 Oct 27 14:02 tty02 crw-rw-rw- 1 root system 35, 3 Oct 27 14:02 tty03
The
/etc/remote
file must be modified to add new access line definitions
for the PCMCIA modem cards to be used. If you have a 28.8kpb modem card
and will be using the full speed, the baud rate (br) in the
/etc/remote
file should be set to 38400.
For example, add the following line to /etc/remote file:
line2:dv=/dev/tty02:br#38400:pa=none:
Note that in the above line, "line2" can be any name you determine to be used with the "tip" command later on.
Once the PCMCIA modem card is inserted correctly and the system configures the card, the card can be used the same as any other modem devices.
To use a PCMCIA modem card, insert the card to one of the PC Card slots in the PCMCIA adapter. Depending on the adapter type, there may be 2 front access card slots or one front access and one rear access card slot. When you insert the card into the slot 0, you should see the following message on the console terminal (or the Console Log window of the graphics head).
# PCMCIA socket 0: card manufacturer: MEGAHERTZ product name: XJ2288 Configured: serial unit 2, type=16550A ace2 at pcmcia0
This example used MegaHertz XJ2288 fax/modem card.
When a modem card is inserted, an error message such as the following example may appear on the Console Log window:
socket 0: card manufacturer: MEGAHERTZ, unknown modem card inserted
Using generic modem driver for this PC Card.
PCMCIA socket 0: card manufacturer: MEGAHERTZ
product name: XJ1144
socket 0: Couldn't find usable config. for this card. Please eject this PC Card.
The error message:
Couldn't find usable config. for this card.
is generated if the card requires I/O resources that are already in use by other components in the system. If this error message is seen, the card should be ejected since it is not configured.
One possible way to correct this situation is to remove some other ISA/EISA devices in the system and reboot the system, thus freeing I/O resources that may be required by this card and trying the card insertion again.
Once you are finished using the modem card, push the button next to the card slot to eject the card previously inserted. You should see the following message on the console terminal or console Log window.
# stray interrupt on unit=2, intr_id=0 PCMCIA socket 0: PC Card removed
Note that you may or may not see the "stray interrupt..." message above when you eject the card. The message is generated by the serial-line driver if the PC Card generated an interrupt when the card got ejected.
Dynamic Device Recognition (DDR) is a framework for describing the operating parameters and characteristics of SCSI devices to the SCSI CAM I/O subsystem. You can use DDR to include new and changed SCSI devices into your environment without having to reboot the operating system. You do not disrupt user services and processes, as happens with static methods of device recognition.
Beginning with DIGITAL UNIX Version 4.0, DDR is preferred over the
current, static method for
recognizing
SCSI devices.
The current, static method, as described in
System Administration,
is to edit SCSI device customizations into the
/sys/data/cam_data.c
data file,
reconfigure the kernel, and shut down and
reboot the operating system.
Note
Support for the static method of recognizing SCSI devices will be retired in a future release of DIGITAL UNIX .
DIGITAL UNIX Version 4.0 supports both methods of recognizing SCSI devices. Both methods can be employed on the same system, with the restriction that the devices described by each method are exclusive to that method (nothing is doubly-defined).
The information DDR provides about SCSI devices is needed by SCSI
drivers. You can supply this information using DDR when you add new
SCSI devices to the system, or you can use the
/sys/data/cam_data.c
data file
and static configuration methods.
The
information
provided by DDR and the
cam_data.c
file have the same objectives.
When compared to the static method of providing SCSI device
information,
DDR minimizes the amount of information that
is supplied by the
device driver or subsystem to the operating system and maximizes
the amount of information that is supplied by the device itself or by
defaults specified in the DDR databases.
You can also use DDR capabilities to convert customizations in the
cam_data.c
file to information in the DDR
/etc/ddr.dbase
text database.
For more information about DDR, see
System Administration,
ddr_config(8),
and
ddr.dbase(4).