OpenVMS VMS721_UPDATE-V0100 Update Alpha V7.2-1 ECO Summary
TITLE: OpenVMS VMS721_UPDATE-V0100 Update Alpha V7.2-1 ECO Summary
Modification Date: 18-APR-2000
Modification Type: Documentation Modification:
Superseded kit information updated.
Bugcheck caution added.
NOTE: An OpenVMS saveset or PCSI installation file is stored
on the Internet in a self-expanding compressed file.
For OpenVMS savesets, the name of the compressed saveset
file will be kit_name.a-dcx_vaxexe for OpenVMS VAX or
kit_name.a-dcx_axpexe for OpenVMS Alpha. Once the OpenVMS
saveset is copied to your system, expand the compressed
saveset by typing RUN kitname.dcx_vaxexe or kitname.dcx_alpexe.
For PCSI files, once the PCSI file is copied to your system,
rename the PCSI file to kitname-dcx_axpexe.pcsi, then it can
be expanded by typing RUN kitname-dcx_axpexe.pcsi. The resultant
file will be the PCSI installation file which can be used to install
the ECO.
Copyright (c) Compaq Computer Corporation 2000. All rights reserved.
OP/SYS: OpenVMS Alpha V7.2-1
COMPONENT: System
SOURCE: Compaq Computer Corporation
ECO INFORMATION:
ECO Kit Name: VMS721_UPDATE-V0100
DEC-AXPVMS-VMS721_UPDATE-V0100--4.PCSI
ECO Kits Superseded by This ECO Kit: VMS721_DS20E_ES40-V0100
VMS721_F11X-V0100
VMS721_FIBRECHAN-V0200
VMS721_GRAPHICS-V0300
VMS721_PTHREAD-V0100
VMS721_SHADOWING-V0100
VMS721_SYS-V0200
ECO Kit Approximate Size: 95,632 Blocks
Kit Applies To: OpenVMS Alpha V7.2-1
System/Cluster Reboot Necessary: Yes
Rolling Re-boot Supported: Yes
Installation Rating: INSTALL_1
1 - To be installed on all systems running
the listed version(s) of OpenVMS.
Kit Dependencies:
The following remedial kit(s) must be installed BEFORE
installation of this kit:
None
In order to receive all the corrections listed in this
kit, the following remedial kits should also be installed:
None
*******************************************************************
* CAUTION *
*******************************************************************
* For the corrections in this ECO to take effect, the system must *
* be rebooted. If this ECO is installed but the system is NOT *
* rebooted, the system may crash with a BADVECTOR bugcheck when *
* using DEBUG. *
*******************************************************************
ECO KIT SUMMARY:
An ECO kit exists for System Components on OpenVMS Alpha V7.2-1. This
kit addresses the following problems:
Problems Addressed in VMS721_UPDATE-V0100 Rated as "1":
o When a copy operation that interrupts a merge operation
terminates, it finds that there are no members marked for
merge and the thread crashes the system with a SHADDETINCON
bugcheck.
Images Affected:
- [SYS$LDR]SYS$SHDRIVER.EXE
- [SYSEXE]SHADOW_SERVER.EXE
o A system crash occurred with a SHADDETICON bugcheck.
Images Affected:
- [SYS$LDR]SYS$SHDRIVER.EXE
- [SYSEXE]SHADOW_SERVER.EXE
o A system may crash with a KRNLSTAKNV bugcheck during heavy
disk I/O activity, such as BACKUP.
Images Affected:
- [SYS$LDR]IO_ROUTINES.EXE
- [SYS$LDR]IO_ROUTINES_MON.EXE
o An attempt to close a timed out file in mount verify will fail
to close, causing FILCNTNONZ bugchecks during process rundown.
The symptom is that the DEACCESS QIO returns with an error
status, typically VOLINV, as well as CCB$L_WIND containing
zero, which means that the file has been successfully
DECACCESSed.
Images Affected: [SYS$LDR]IO_ROUTINES.EXE
o The external priority value SCH$CHANGE_CUR_PRIORITY is not
always used to calculate the correct entry in the CPU's
scheduling array. Instead, the internal priority value will
be used if another CPU is executing at the same process
priority as the current CPU that is executing the
SCH$CHANGE_CUR_PRIORITY routine. Using the internal process
priority, could result in data corruption since the calculated
address of an entry in the CPU's scheduling array will most
likely be beyond the end of this array and also beyond the end
of the per-CPU database structure. So, whatever data lies
after this could get corrupted.
Images Affected:
- [SYS$LDR]PROCESS_MANAGEMENT.EXE
- [SYS$LDR]PROCESS_MANAGEMENT_MON.EXE
o During SHUTDOWN of a Galaxy instance, if the shared memory LAN
driver is active, the system could crash with an INVEXCEPTN.
Images Affected:
- [SYS$LDR]SYS$EBDRIVER.EXE
o An INVEXCEPTN crash can occur.
Images Affected:
- [SYS$LDR]SYSDEVICE.EXE
o Enable_lun was not functioning properly, resulting in possible
crashes in EXE$INSTIMQ.
Images Affected:
- [SYS$LDR]SYS$PKQDRIVER.EXE
o A SCSI$C_RESERVATION_CONFLICT with a data underrun condition
causes a bugcheck.
Images Affected:
- [SYS$LDR]:SYS$PKQDRIVER.EXE
o A bad disk in a shadow set is not thrown out of the set, but
instead causes the shadow set to hang, which may hang the
system.
Images Affected:
- [SYS$LDR]SYS$PKCDRIVER.EXE
- [SYS$LDR]SYS$PKEDRIVER.EXE
- [SYS$LDR]SYS$PKQDRIVER.EXE
- [SYS$LDR]SYS$PKSDRIVER.EXE
- [SYS$LDR]SYS$PKTDRIVER.EXE
- [SYS$LDR]SYS$PKWDRIVER.EXE
- [SYS$LDR]SYS$PKZDRIVER.EXE
o During SHUTDOWN of a Galaxy instance, if other interconnects
are in use besides SMCI, there exists the possibility of other
Galaxy instances crashing with a CLUEXIT.
Images Affected:
- [SYS$LDR]SYS$PBDRIVER.EXE
o An ACCVIO crash could occur when a SMCI port is shutdown.
Images Affected:
- [SYS$LDR]SYS$PBDRIVER.EXE
o Running ERRFMT interactively can cause the system to crash
with a SSRVEXCEPT bugcheck.
Images Affected:
- [SYSEXE]ERRFMT.EXE
o The system can crash when trying to create a directory or
write a file in a large InfoServer partition.
Images Affected:
- [SYS$LDR]SYS$DADDRIVER.EXE
o A mismatch between the actual and claimed map register crashes
the system.
Images Affected:
- [SYS$LDR]SYS$DRDRIVER.EXE
o Several of the Galaxy locking services can crash the system
with an access violation in the kernel mode if used in a
"share nothing" configuration.
Images Affected:
- [SYS$LDR]SYS$GALAXY.EXE
o If a process working set is reduced such that the working set
only contains locked pages, and the process then pagefaults,
the system could bugcheck in MMG$PAGEFAULT with "FREWSLX,
Images Affected:
- [SYS$LDR]SYS$VM.EXE
o A situation in which PT regions share VA space can cause a
crash.
Images Affected:
- [SYS$LDR]SYS$VM.EXE
o The UCB$L_RECORD field, which should = 0 after a successful
rewind, contains erroneous values. These values are being
returned by the controller in the end message field
MSCP$L_POSITION.MOUNTVER/TAPE_VALIDATION. The system then
uses these values as an index and the system ACCVIO crashes.
Images Affected:
- [SYS$LDR]SYS$TUDRIVER.EXE
TMSCP resources, it is best to fix this in TUDRIVER.
o During a system boot with SYSTEM_CHECK set to 0 and with XFC
(Extended File Cache) loaded (VCC_FLAGS = 2), a system could
crash.
Images Affected:
- [SYS$LDR]MESSAGE_ROUTINES.EXE
o A reboot on a DPWS (Digital Personal Workstation) with an ELSA
card installed will crash. The system must be shut down, a
console init must be performed and then the system must be
booted.
Images Affected:
- [SYS$LDR]SYS$GZDRIVER.EXE
- [SYS$LDR]SYS$GZVDRIVER.EXE
o Timer routine LAST$TICK causes a crash.
Images Affected:
- [SYS$LDR]SYS$LASTDRIVER.EXE
o When the AMDS console is writing to RMDRIVER, a system can
crash with an PGFIPLHI bugcheck.
Images Affected:
- [SYS$LDR]SYS$SMDRIVER.EXE
o Systems with heavy system pagefaulting could incur a PGFIPLHI,
"Pagefault with IPL too high," bugcheck.
Images Affected:
- [SYSLDR]SECURITY.EXE
- [SYSLDR]SECURITY_MON.EXE
o Two class schedulers running on the same system could cause a
crash when they both tried to insert the same TQE onto the
system-wide TQE queue multiple times.
Images Affected:
- [SYS$LDR]CLASS_SCHEDULER.EXE
o An internal TTDRIVER bug in CTRL/R processing can cause a
corruption of non-paged pool, resulting in a variety of system
bugcheck crashes, such as INVEXCEPTN, SSRVEXCEPT, LOCKMGRERR
or others.
Images Affected:
- [SYS$LDR]SYS$TTDRIVER.EXE
o In rare cases, a system may crash with a SYSTRMERR bugcheck
during a boot, if verbose boot messages are enabled. Setting
bit 16 or 17 in the console's OS_FLAGS parameter (boot -fl
x,10000 or x,20000 where x is the system's root directory)
enables verbose boot messages.
Images Affected:
- [SYSEXE]STACONFIG.EXE
o The bug in the printf() "engine" introduced in the
ALPACRT07_071 kit has been fixed. The bug was that when the
minimal field width or precision specified in the format
directive was greater than 1024, the "engine" could write
beyond the end of allocated internal conversion buffer. It
could happen only in X_FLOAT floating point mode
(/L_DOUBLE_SIZE=128, which is the default on Alpha). The bug
affected all functions from the printf() family of functions.
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o A system crash could occur as the result of a delay in the
downward adjustment of the SP (Stack Pointer), which could
corrupt the stack.
Images Affected:
- [SYS$LDR]SYS$DKDRIVER.EXE
- [SYS$LDR]SYS$GKDRIVER.EXE
- [SYS$LDR]SYS$MKDRIVER.EXE
o The PCI FDDI adapter halts with a fatal hardware error. When
the driver attempts to restart it, the driver detects either a
parity error being reported by the DEFPA, or a command
failure, which prevents the driver from getting the DEFPA
restarted successfully. As a result, the DEFPA is unusable
until the system is rebooted.
Images Affected:
- [SYS$LDR]SYS$FWDRIVER.EXE
- [SYS$LDR]SYS$FRDRIVER.EXE
- [SYS$LDR]SYS$FADRIVER.EXE
Problems Addressed in VMS721_UPDATE-V0100 Rated as "3":
o INSTALL REPLACE of an image installed resident, when there
sufficient granularity hint region space available, fails to
correctly load resident code (and/or data). Any subsequent
use of the replaced image is very likely to fail, Access
violations (ACCVIO) are likely to occur.
Images Affected:
- [SYSEXE]INSTALL.EXE
o The getenv() function has been corrected to check for a
Command Line Interpreter (CLI) symbol only when called in user
access mode. While DEC C RTL is a user-mode run time library,
a decision was made to correct getenv() in order to make the
library more robust. The reason for this change is because
calling the CLI from within exec or kernel mode may result in
an access violation.
With this fix, when called from inner access mode, the
getenv() function will search only the C environment list and
OpenVMS logical names.
This fix does not mean, however, that the user-mode only
restriction was lifted. Calling the DEC C RTL functions,
including getenv(), in any access mode other than in user-mode
is still un-supported.
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o The strptime() function has been modified to comply with
X/Open CAE Specification System Interfaces and Headers Issue 5
(commonly known as XPG5). In XPG5, how the strptime()
function processes the "%y" directive was changed for a
two-digit year within the century, if no century is specified.
According to XPG5, for the "%y" directive, when a century is
not otherwise specified, values in the range 69-99 refer to
years in the twentieth century (1969 to 1999 inclusive), while
values in the range 00-68 refer to years in the twenty-first
century (2000 to 2068 inclusive). Essentially, for the "%y"
directive, strptime() became a "pivoting" function with 69
being a pivoting year.
Before this change, the strptime() function was always
interpreting a two-digit year with no century as a year within
the twentieth century.
With the current ECO kit, XPG5-compliant strptime() becomes a
default strptime() function in the DEC C RTL. However, for
compatibility reasons, the previous "non-pivoting"
XPG4-compliant strptime() function was retained.
The "pivoting" is controlled by the DECC$XPG4_STRPTIME logical
name. Defining this logical name prior to invoking the
application (any equivalence string) will cause the DEC C RTL
to use the "non-pivoting" flavor of strptime(). Also,
"non-pivoting" strptime() can be called directly as the
decc$strptime_xpg4() function.
While there is an easy way to disable "pivoting" by just
defining a logical name, you should be aware that if the
application ceases to work correctly due to the change in the
strptime() function, it almost certainly means that there is a
Y2K problem in the application. Y2K-safe applications should
not be affected by the change in strptime().
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o The stat() function has been corrected to process file
specifications such as "foo:[000000]", where foo is defined as
a concealed device like the following:
$ define/trans=(conc) foo device:[bar.]
Prior to this fix, the stat() function would fail for such a
file specification with errno set to ENOENT (No such file or
directory).
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o Both flavors of the system() function were corrected to ensure
that the function always returns the correct status.
The POSIX-compliant system() function has been corrected to
always return a status corresponding to the signal, which
terminated the child process if the child process was
terminated due to receiving a signal.
Prior to this fix, there was a small timing window when the
function could return the status corresponding to the child
process completion code, even in the case when the child
process was terminated by a signal.
The Non-POSIX system() function has been corrected to always
return the OpenVMS completion code of the child process.
Prior to this fix, there was a small timing window when the
function could return the status from LIB$SPAWN(), instead of
the completion code of the child process.
None of these problems were ever reported by a customer.
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o The opendir() and readdir() functions have been corrected to
process directories on a remote node specified using the
following DECnet file specification:
nodename"username password"::device:[directory]
Prior to this fix, opendir() could fail with errno set to
ENOTDIR (Not a directory). In the case when opendir() could
successfully open the directory, the readdir() function could
return incorrect file names.
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o The times() and clock() functions have been modified to be
both thread-safe and AST-reentrant.
Prior to this change, these functions belonged to the class of
functions which are thread safe, but not AST-reentrant (see
section 1.7.2 Multithread Restrictions in the DEC C RTL
Reference Manual).
The change in the reentrancy status of the times() and clock()
functions will be reflected in the next release of the Manual.
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o The stat() function no longer fails due to the failure of the
mktime() function.
It was reported that the stat() function fails for a file
created on April 4th 1999 at 2:04 a.m. EST, which is the
"missing hour" in the transition from standard to daylight
saving time in the Eastern time zone.
The problem was due to the failure of the mktime() function to
process the "missing hour" (stat() calls mktime() to fill
st_?time members of the stat structure with the UTC time).
While the mktime() function was not fixed, the stat() function
was corrected to handle the failure of mktime() and set
st_?time fields correctly, even if mktime() fails.
Note, that, strictly speaking, "missing hour" is an invalid
input for mktime(): there is no such time as "April 4th 1999
2:04 a.m.", for example, in the Eastern time zone. Given the
possibility of calling mktime() with an invalid time, the
stat() function probably should not have used mktime() in the
first place or at least should have been prepared for a
mktime() failure.
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o The CRTL I/O subsystem was enhanced to lift the eight nested
directory levels restriction for an ODS-5 device. This change
affects such DEC C RTL functions as access, mkdir, opendir,
rmdir, and stat.
The following program example shows different OpenVMS and
Unix-style deep directory specifications.
#include
#include
#include
#include
#include
#include
#define DIM(_s) (sizeof(_s) / sizeof(_s[0]))
/*
** we assume presence of root$ concealed device:
**
** define/NoLog/Tran=(Conceal,Terminal) root$ node$dka0:[000000.]
**
*/
static char *filename[] = {
"node$dka0:[w.x.y.z.a.b.c.d.e.f.g.h.i.j]", /* (1) */
"node$dka0:[w.x.y.z.a.b.c.d.e.f.g.h.i]j.dir", /* (2) */
"root$/w/x/y/z/a/b/c/d/e/f/g/h/i/j", /* (3) */
};
main()
{
struct stat s;
int i;
char *name;
for(i=0; i < DIM(filename); ++i) {
name = filename[i];
if ( mkdir(name,0) == -1 )
printf("%d? mkdir(): %s\n", i+1, strerror(errno));
if ( stat(name,&s) )
printf("%d? stat(): %s\n", i+1, strerror(errno));
else if ( !S_ISDIR(s.st_mode) )
printf("%d? stat(): %s\n", i+1, "not a directory???");
if ( !opendir(name) )
printf("%d? opendir(): %s\n", i+1, strerror(errno));
if ( access(name,F_OK) == -1 )
printf("%d? access(): %s\n", i+1, strerror(errno));
if ( rmdir(name) == -1 )
printf("%d? rmdir(): %s\n", i+1, strerror(errno));
}
}
Please note that the DEC C RTL still does not support mixed
Unix/OpenVMS-style filenames coded using extended ODS-5
syntax, such as "doc/foo.bar.bar", "~foo^.bar" and
"./tmp/foo.bar.b^_ar".
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o The select() function has been corrected to return a failure
status if either an invalid file descriptor or file descriptor
not associated with a socket is found in one of the specified
file descriptor sets. In the case of an invalid file
descriptor, the select() function sets errno to EBADF. In
case of a file descriptor not associated with a socket, the
function sets errno to ENOTSOCK.
Failure with errno set to EBADF is the standard requirement
for the select() function. Failure with errno set to ENOTSOCK
occurs because the select() function can currently operate
only on sockets.
Prior to this fix, the function was setting errno as described
above, but otherwise was ignoring invalid file descriptors and
file descriptors not associated with sockets.
The old behavior can be requested by defining the logical name
DECC$SELECT_IGNORES_INVALID_FD prior to invoking the
application (any equivalence string).
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o The problem commonly known as the "Y2038 bug" has been fixed.
The C run-time library time functions, which access the
current time using the OpenVMS time functions, have been
corrected to handle times after 19-Jan-2038 3:14:07. Data
type time_t is defined on OpenVMS as an unsigned 32-bit
integer. However, some of the functions which retrieve
current time from the system and convert it to time_t format
were using signed arithmetic.
This fix affects the functions:
o decc$fix_time()
o time()
o ftime()
o gettimeofday()
o getclock()
With this fix, DEC C RTL time functions will handle times
successfully until 07-Feb-2106 06:28:15.
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o The printf() family of functions have been enhanced to perform
better when the calling application has exhausted available
memory.
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o The file access functions have been corrected to process file
names on UNIX systems over DECnet. The file name for a file
on a UNIX system accessed over DECnet was being converted to
uppercase, even when the name was enclosed in double quotes.
This problem was introduced on OpenVMS V7.2 and also affects
OpenVMS V7.1-2 and OpenVMS V7.2-1. File names of the form
node::"dir/name" are now processed correctly.
Images Affected:
- [SYSLIB]DECC$SHR.EXE
- [SYSLIB]CRTL.OBJ
- [SYSLIB]CRTLMSGDEF.OBJ
o When installing FORTRAN and FORRTL products, the system hangs
if both products are selected for installation at the same
time.
Images Affected:
- [SYSLIB]PCSI$SHR.EXE
- [SYSEXE]PCSI$MAIN.EXE
o If the following PCSI steps are taken:
1. Install a FULL type product kit
2. Patch the FULL product with a PATCH type kit(s)
3. Install a PARTIAL type software product on top of the FULL
kit
then, two versions of the same FULL product are created and
patches are not transferred from the FULL product to the
PARTIAL product. This would eventually lead to an
un-repairable PCSI database corruption.
Images Affected:
- [SYSEXE]PCSI$MAIN.EXE
- [SYSLIB]PCSI$SHR.EXE
o The following duplicate error messages reported by the PCSI
utility display during a product upgrade:
%PCSI-E-INSVOLSPC, insufficient space on volume DISK$ALPCSI_SYS
-PCSI-I-VOLSPC, 159261 required; 124344 available; -34917 net
Terminating is strongly recommended.
Do you want to terminate? [YES] no
%PCSI-E-INSVOLSPC, insufficient space on volume DISK$ALPCSI_SYS
-PCSI-I-VOLSPC, 159261 required; 124344 available; -34917 net
Terminating is strongly recommended.
Do you want to terminate? [YES] no
Images Affected:
- [SYSLIB]PCSI$SHR.EXE
o The reinstallation of a PARTIAL kit causes an ACCVIO. This
occurs when certain data is passed from one instance of the
PARTIAL product to another instance of the same product.
After the transfer, pointers to the originating data
structures are cleared. These pointers will be needed later.
When the PARTIAL kits are different, re-installation of
PARTIAL kit will not clear pointers.
Images Affected:
- [SYSLIB]PCSI$SHR.EXE
o An OpenVMS upgrade would generate an unexpected prompt to
terminate the operation.
Images Affected:
- [SYSLIB]PCSI$SHR.EXE
- [SYSEXE]PCSI$MAIN.EXE
o When re-installing a product that has a library module in its
Product Description File (PDF), the PCSI utility terminates
with an ACCVIO error.
Images Affected:
- [SYSLIB]PCSI$SHR.EXE
- [SYSEXE]PCSI$MAIN.EXE
o Changed the behavior of file conflict resolution between a
file from the kit being installed and a file with the same
name that has already been installed. Previously, when two
files had the same generation number, the PCSI utilty would
retain the file already on the system disk. Now when there is
a tie, the file from the kit being installed replaces the file
previously installed.
Images Affected:
- [SYSLIB]PCSI$SHR.EXE
- [SYSEXE]PCSI$MAIN.EXE
o When a PRODUCT RECONFIGURE command is executed and the
reconfigured product has two or more patches applied to it,
the PCSI utility incorrectly marks the patches as removed.
This inconsistency can cause an ACCVIO on subsequent
operations.
Images Affected:
- [SYSLIB]PCSI$SHR.EXE
o When an OpenVMS partial kit is installed to upgrade the full
version of OpenVMS, new files supplied by the partial kit that
should replace the ones provided by the full OpenVMS version
are ignored by the PCSI utility.
Images Affected:
- [SYSLIB]PCSI$SHR.EXE
o The PCSI utility returns the %PCSI-F-ASSERTFAIL error
message and aborts when upgrading the OpenVMS operating system
with a partial kit.
Images Affected:
- [SYSLIB]PCSI$SHR.EXE
o When a FULL product kit replaces a file provided by another
product, or a PATCH product kit replaces a file provided by
its parent product or another patch, use of the IMAGE LIBRARY
option of the FILE statement in the PDF (Product Description
File) must be used consistently. For example, if one product
provides a file and places it in SYS$LIBRARY:IMAGELIB.OLB, but
another product replaces the file but does not specify that it
should also be placed in image library, then the installation
will fail with a %PCSI-E-PARUDF error message.
Images Affected:
- [SYSLIB]PCSI$SHR.EXE
o Installing OpenVMS patch kits out of order results in a
%SYSTEM-F-ACCVIO access violation. The proper order, assuming
OpenVMS is already loaded, is:
1. Install the latest patch first, then any additional
patches in descending order, based upon Release Version,
e.g., V5.1, V4.2, V3.4, etc.
2. Install the patch UPDATE V1.0, then UPDATE02 V1.0, and
then re-install patch UPDATE V1.0.
Images Affected:
- [SYSLIB]PCSI$SHR.EXE
o Correct the resultant directory string size in NAML for remote
file specifications.
The size of the resultant directory string in a NAML block
structure was being set incorrectly when parsing remote file
specifications.
Images Affected:
- [SYS$LDR]RMS.EXE
o Fix to wildcard '%' when % literal also in next subdirectory
level.
When doing an RMS $PARSE (without SYNTAX-ONLY, so the
existence of the directory is confirmed) of a file
specification with a subdirectory reference at level n with a
wildcard percent sign (%) and a subdirectory name at level n+1
with a literal percent sign (^%), RMS will return a syntax
error, even though the specification is valid (and the
directory exists).
This problem only occurs with an ODS-5 disk (which allows
names with literal percent signs).
Images Affected:
- [SYS$LDR]RMS.EXE
o Correction for processes exiting with RMS IORNDN non-fatal
bugcheck.
Processes may disappear with RMS IORNDN non-fatal bugchecks
when an EXIT is requested by an Executive-mode application
(such as ACMS). This is a very small timing window, so
processes with a large number of files increases the
probability of the problem occurring.
If the SYSGEN parameter BUGCHECKFATAL is not enabled, then the
process will be terminated; if it is enabled, then the system
will crash with a RMSBUG (R2=FFFFFFF0, IORNDN) bugcheck.
Images Affected:
- [SYS$LDR]RMS.EXE
o Fix to allow mixed (% and ?) wildcards in directory or
filename specification.
If within one filespec a mixture of '%' and '?' is used as the
one-character wildcard in a directory or filename
specification, a $PARSE or $SEARCH fails with 'file not
found', even though there is a subdirectory or file name on
disk that matches the wildcard specification. For this
problem to occur using the DCL interface with a file located
on an ODS-2 disk, the PARSE_STYLE for the process would need
to be set to EXTENDED.
Images Affected:
- [SYS$LDR]RMS.EXE
o The CONVERT utility may return an access violation and/or
SORT_ON errors when it is repeatedly invoked from within an
application utilizing the callable interface. Additionally,
an invalid file structure may be created when the callable
interface is invoked repeatedly from within an application.
Images Affected:
- [SYSLIB]CONVSHR.EXE
o Attempts to CONVERT a prologue 3 indexed file with greater
than 3 keys (key 3 and above) and with the primary key being
segmented with the segments not in ascending order, results
in invalid key structures being generated for key 3
and above.
1. Using the callable interface to convert multiple files,
results in a SORT_ON error if any previous file contains
no records and has at least one alternate key defined.
2. Using the /SECONDARY qualifier with values greater than
eight results in invalid alternate key structures being
generated.
Images Affected:
- [SYSLIB]CONVSHR.EXE
o CONVERT/RECLAIM reports the following error during image
rundown:
"%RMS-F-ISI, invalid internal stream identifier (ISI) value"
Despite the reported error, the CONVERT/RECLAIM completes
successfully.
Images Affected:
- [SYSLIB]CONVSHR.EXE
o Fix %CONVERT-I-SEQ errors converting sequential file to
indexed.
Attempts to convert a sequential file to an indexed format may
report %CONVERT-I-SEQ errors despite convert's invocation of
the SORT utility. These errors may be reported if any of the
input file's records are shorter than the primary key's
highest segment.
Images Affected:
- [SYSLIB]CONVSHR.EXE
o The IO User's Guide states the the IO$_FLUSH function is
supported. However, when and IO$_FLUSH is called, an "illegal
IO function" error message is returned.
Images Affected:
- [SYS$LDR]]SYS$MKDRIVER.EXE
o Fatal IVDENS (invalid density) errors can occur during simple
operations, such as INIT, on a TZ88 tape drive.
Images Affected:
- [SYS$LDR]SYS$MKDRIVER.EXE
o If a write protected disk is dismounted and un-write
protected, when the disk is remounted, it is still software
write protected.
Images Affected:
- [SYS$LDR]SYS$DKDRIVER.EXE
o Appendix B of the OpenVMS V7.2 I/O User's Reference Manual
states that the following values are ignored when
S2DGB$V_TAGGED_REQ is 1:
S2DGV$L_32PHSTMO
S2DGV$L_64PHSTMO
S2DGV$L_32DSCTMO
S2DGV$L_64DSCTMO
S2DGB$V_DISCPRIV
Although not documented, the following values are also
ignored:
2DGV$L_32PADCNT
S2DGV$L_64PADCNT
The documentation also stated that ports which do not support
tagged command queuing always behave as if S2DGB$V_TAGGED_REQ
is 0. This applies to the behavior of the ports and has no
effect on whether the above set of parameters gets ignored.
The reason these values are ignored when tagged command
queuing is in use is that they can effect other commands to
the connection until the IO$_DIAGNOSE command completes. (The
timeout values are used as defaults for all commands to the
connection for the duration of the command.)
Images Affected:
- [SYS$LDR]SYS$DKDRIVER.EXE
- [SYS$LDR]SYS$MKDRIVER.EXE
- [SYS$LDR]SYS$GKDRIVER.EXE
o A generic scsi device may cause command timeout errors in the
error log file.
Images Affected:
- [SYS$LDR]SYS$DKDRIVER.EXE
- [SYS$LDR]SYS$GKDRIVER.EXE
- [SYS$LDR]SYS$MKDRIVER.EXE
- [SYS$LDR]SYS$PKEDRIVER.EXE
o Processes get stuck in RWNPG state waiting to write to MBA1.
An ANALYZE/SYSTEM command shows that the JOB_CONTROL process
has no QIO against MBA1. As a result, the SHOW/QUEUE/MANAGER
command hangs and new users cannot log into the system.
Images Affected:
- [SYSEXE]JBC$JOB_CONTROL.EXE
o If IMAGE ACCOUNTING is enabled, an IMAGE termination record
may be written to ACCOUNTNG.DAT as a USER record.
Images Affected:
- [SYSEXE]JBC$JOB_CONTROL.EXE
o The DCL SHOW QUEUE or SHOW ENTRY commands display batch jobs
executing or aborting when there are no corresponding
processes for the jobs active on the system.
The accounting record shows the jobs as terminated.
Images Affected:
- [SYSEXE]JBC$JOB_CONTROL.EXE
o When a queue is deleted, the queue manager can issue a SIGNAL
stop and crash with the error message "%QMAN-F-BADLOGIC,
internal logic error detected."
Images Affected:
- [SYSEXE]QMAN$QUEUE_MANAGER.EXE
- [SYSEXE]JBC$JOB_CONTROL.EXE
o The job controller process can hang in an LEF state due to an
event flag zero overlap. As a result, queues and jobs remain
in the starting state.
Images Affected:
- [SYSEXE]JBC$JOB_CONTROL.EXE
- [SYSEXE]QMAN$QUEUE_MANAGER.EXE
o The following two problems occur:
o Init/density=(1600,6250) commands do not honor the
requested density.
o An INIT/MEDIA=compaction command does not change to
compacted mode.
The problems exist when the configuration includes an OpenVMS
V7.1 client and an OpenVMS V7.2 server utilizing an
OpenVMS-TMSCP served MK-based tape drive. The drives which
have exhibited this problem include TZ8X, TSZ07, TLZX or any
SCSI-based drive that supports data compaction.
Images Affected:
- [SYS$LDR]TMSCP.EXE
o The device type assigned to a tape drive can be inconsistent
between a client and the server. When a local SCSI tape drive
is configured on a server and assigned a device type, such as
"DEC XXX", where XXX is equal to TLZ10 or any other type of
SCSI tape drive, when the client configures the device, a
device type "GENERIC SCSI TAPE" is assigned for the same local
SCSI tape drive.
Images Affected:
- [SYS$LDR]TMSCP.EXE
o MSCP HRB structures are overwritten by other driver modules.
The structures cause crashes in other modules when they
attempt to use a MSCP HRB as another data structure.
Images Affected:
- [SYS$LDR]MSCP.EXE
o Commands which specify an invalid density execute without
indicating an error condition.
Images Affected:
- [SYS$LDR]SYS$TUDRIVER.EXE
o An INIT/MED=NOCOMP command to an HSJ-based DLT (TZXX) drive
does not go into non-compacted mode.
Images Affected:
- [SYS$LDR]SYS$TUDRIVER.EXE
o SHOW DEVICES shows zero % merged status, even though the
shadow set status does not indicate that a merge is required.
In addition, the shadow set shows that it has merge members,
even though the SHAD$W_STATUS does not indicate that
condition.
Images Affected:
- [SYS$LDR]SYS$SHDRIVER.EXE
- [SYSEXE]SHADOW_SERVER.EXE
o The various reasons for generating SS$_INCSHAMEM are not
traced.
Images Affected:
- [SYS$LDR]SYS$SHDRIVER.EXE
- [SYSEXE]SHADOW_SERVER.EXE
o The DQDRIVER's DMA windows in PCIbus space are shifted out of
natural alignment, if the DRDRIVER for the KZPAC Mylex RAID
controller configures before the DQDRIVER. As a result, the
DQDRIVER is not configured properly and adversely affects
other structures in memory.
Images Affected:
- [SYS$LDR]SYS$DQDRIVER.EXE
o The Cypress chip does not reset its DMA pointer if the transfer
is an exact multiple of 8KB.
Images Affected:
- [SYS$LDR]SYS$DQDRIVER.EXE
o When the following occurs:
1. A username exists in both the SYSUAF.DAT and the SAM
database
2. The IGNORE_EXTAUTH bit is set in the SECURITY_POLICY
system parameter
An $ACM call that is targeted to the NT ACME for a username
that is recognized by both OpenVMS and NT fails.
Images Affected:
- [SYSLIB]VMS$VMS_ACMESHR.EXE
o Specifying a complete filename with a relative (negative) file
version is not correctly processed by BACKUP. No files are
selected.
Images Affected:
- [SYSEXE]BACKUP.EXE
- [SYSLIB]BACKUPSHR.EXE
o BACKUP/JOURNAL/SELECT may fail with an %RMS-F-SYN file
specification syntax error. The failure occurs when a journal
entry for a header only file is encountered in the journal
file, e.g., [SYS0.SYSEXE]SWAPFILE.SYS;1.
Images Affected:
- [SYSEXE]BACKUP.EXE
- [SYSLIB]BACKUPSHR.EXE
o When SYSMAN is used to access another node in the same
cluster, and the access fails due to a bad password, the
intrusion record is created using a 12 byte space padded field
for the USERNAME, such as,
"SYSTEM::USER "
When the user does a SHOW INTRUSIONS, they do not see the
intrusion record as a 12 byte space padded field but rather as
just "SYSTEM::USER". Trying to delete the intrusion record
with a DEL/INTRUSION "SYSTEM::USER" does not work. In order
to delete the record, the 12 byte space padded field format
must be entered:
DEL/INTRUSION "SYSTEM::USER "
This correction eliminates the need to use a 12 byte space
padded field to delete the intrusion record.
Images Affected:
- [SYSEXE]SMISERVER.EXE
o Shutting down a node which boots from a V7.1-2, or later,
shadowed system disk results in a shadow merge operation,
which causes performance degradation..
Images Affected:
- [SYSEXE]SMISERVER.EXE
- [SYSEXE]SHUTDOWN.COM
o The number of pagelets that can be specified for SYSGEN
parameters, such GBLPAGES, is limited.
Images Affected:
- [SYSEXE]SMISERVER.EXE
- [SYSEXE]SYSBOOT.EXE
- [SYSEXE]SYSGEN.EXE
- [SYSEXE]SYSMAN.EXE
- [SYSLIB]LIB.R64
- [SYSLIB]SDA$SHARE.EXE
- [SYSUPD]AUTOGEN.COM
o When SDA is run in a batch job, the first page of output is
written to the log, then the "Press return for more" prompt is
output and SDA attempts to read from the batch command file.
In the next record, there is either another SDA command or a
line beginning with a dollar sign. In either case, SDA aborts
the command it was processing.
Images Affected:
- [SYSLIB]SDA$SHARE.EXE
o Some addresses for resident images that are symbolized display
incorrectly or ACCVIO in CLUE.
Images Affected:
- [SYSLIB]SDA$SHARE.EXE
o Several SDA changes were made:
o Make BUGCHECK handle memory holes that start on L2 PTE
boundaries.
o Make an SDA workaround for the lack of access to PTEs or
PFN entries in a dump, which occurs when BUGCHECK does not
correctly dump memory with holes on L2 PTE boundaries.
o Display Floating Point registers correctly with SHOW
PROCESS/REGISTERS.
Images Affected:
- [SYSLIB]SDA$SHARE.EXE
- [SYS$LDR]EXCEPTION.EXE
- [SYS$LDR]EXCEPTION_MON.EXE
o Symbols whose offsets are off the end of the vector are
ignored.
Images Affected:
- [SYSLIB]SDA$SHARE.EXE
o The Registry$server ACCVIOS when Advanced Server tries to
start.
Below is a sample of the REGISTRY$SERVER.log.
%REG-E-DTMUTEXLOCK, DECthreads mutex lock error 22
%REG-E-DTMUTEXUNLOCK, DECthreads mutex unlock error 22
%REG-E-FSOSEGNUMBER, invalid segment number in FSO: 0003038D
00000000 %REG-F-DBACCESS, cannot access registry database
object %CMA-F-EXIT_THREAD, current thread has been requested
to exit %SYSTEM-F-ACCVIO, access violation, reason mask=00,
virtual address 000000031
Images Affected:
- [SYSEXE]REGISTRY$SERVER.EXE
o The registry database cannot be expanded beyond 2 Mg. In
addition, there are deficiencies in the writing of segments
which become more critical when working with large databases.
Images Affected:
- [SYSEXE]REGISTRY$SERVER.EXE
o The NTA_PACK_UNPACK routine required the GID Info Array to be
in the NT Credentials that was being packed and unpacked. The
GID Info Array field should be optional,
Images Affected:
- [SYSLIB]NTA$SECSHR.EXE
- [SYSLIB]NTA$SECSHRP.EXE
- [SYSEXE]REGISTRY$SERVER.EXE
o When an entry is deleted from a paused symbiont queue, an
"unexpected symbiont process" termination record is displayed
in the OPERATOR.LOG and a dump file for the symbiont running
(LATSYM, TELNETSYM, PRTSMB) is produced in the SYS$SYSTEM
directory. An analysis of queue journal files shows each
occurrence of the problem happens after a DELETE/ENTRY of a
print job on a paused queue.
Images Affected:
- [SYSLIB]SMBSRVSHR.EXE
o If a file being printed is a VFC format, and if the print
queue is paused and restarted with "/search = ", then
the print symbiont enters a CPU loop that can consume up to
100% of CPU bandwidth.
Images Affected:
- [SYSLIB]SMBSRVSHR.EXE
o The PROLOSSOR signal may look like %system-?-devforeign.
Several of the stack-walking routines returned status values
that were not being checked.
Images Affected:
- [SYSLIB]DEC$BASRTL.EXE
o MOUNT appears to mount a shadow set that is in "MountVerify
Timeout" state successfully, but in reality, no members are
added and the set remains in the "MountVerify Timeout" state.
A device that is in "MountVerify Timeout" state must be
dismounted before you can successfully re-mount the device.
Images Affected:
- [SYSLIB]MOUNTSHR.EXE
- [SYSEXE]VMOUNT.EXE
o An attempt to MOUNT a disk which has a corrupted INDEX.SYS
file 'hangs' MOUNT. Control-Y is unable to stop the hang and
the operations count on the device rapidly increases.
Images Affected:
- [SYSLIB]MOUNTSHR.EXE
- [SYSEXE]VMOUNT.EXE
o The SYS$MCDRIVER error handler logs a Link-Online De-asserted
error as a Hub Timeout error.
Images Affected:
- [SYS$LDR]SYS$MCDRIVER.EXE
o Hub Timeout errors, MC port re-inits and virtual circuit
closure may occur on Memory Channel 1.0 or 1.5 devices.
Images Affected:
- [SYS$LDR]SYS$MCDRIVER.EXE
o Port Messages are incorrectly being logged as MSCP Tape
Messages.
Images Affected:
- [SYS$LDR]SYS$PBDRIVER.EXE
o PBDRIVER may continually re-init itself every few seconds. On
some re-init paths, PBDRIVER may not properly record the
instance incarnation number.
Images Affected:
- [SYS$LDR]SYS$PBDRIVER.EXE
o The linker aborts with an access violation when linking with
the /SECTION_BINDING qualifier.
Images Affected:
- [SYSEXE]LINK.EXE
o Networks with many devices may experience failures. A $SHOW
DEVICE command returns %SYSTEM-F-VASFUL, "virtual address is
full."
Images Affected:
- [SYSEXE]SHOW.EXE
o Setting PQL_MENQLM above 32767 prevents the special UAF ENQLM
value of 32767 from scaling to "inifinity".
Images Affected:
- [SYSEXE]LOGINOUT.EXE
o Users who login through the DECwindows CDE interface, and who
use external authentication, sometimes experience unexpected
login failures when their mapped username is different than
the username entered during login.
Images Affected:
- [SYSEXE]LOGINOUT.EXE
o In previous releases of OpenVMS, protection of a new FT device
was unconditionally set to S:RWPL,O:RWPL,G,W (no access for
GROUP and WORLD).
This change and a corresponding change in the ALPSYSA04_071
kit modify this behavior. Protection for a new FT device is
now taken from the protection of the "template device" FTA0:.
The protection for FTA0: may be set during the boot process,
(in SYSTARTUP_VMS.COM), or manually, for example
$ SET SECURITY /CLASS=DEVICE-
/PROTECTION=(S:RWLP,O:RWLP,G:RW,W:R) FTA0:
At boot time, the protection on FTA0 is unconditionally set to
S:RWPL,O:RWPL,G,W. An ACL may also be set on FTA0, either
explicitly, or inherited from the SECURITY class TERMINAL
device template.
Images Affected:
- [SYS$LDR]SYS$FTDRIVER.EXE
- [SYS$LDR]IO_ROUTINES.EXE
o Update the conversion of the ODS-1 date field to comply with
the RSX-11 version of date handling.
Images Affected:
- [SYS$LDR]F11BXQP.EXE
o Fix the internal GETTIM function so that it properly formats
the 64 bit internal time to an RSX-11 compatible ASCII string.
Images Affected:
- [SYSLIB]INIT$SHR.EXE
- [SYSEXE]VERIFY.EXE
- [SYS$LDR]F11BXQP.EXE
o Security Server generates OPCOM messages saying it cannot
write to a client mailbox.
Images Affected:
- [SYSLIB]SECURESHRP.EXE
o The stack could overflow when making a call to $FORMAT_ACL.
Images Affected:
- [SYSLIB]SECURESHR.EXE
- [SYSLIB]SECURESHRP.EXE
- [SYSLIB]SECURESHR_JACKET.EXE
o A spurious ACCVIO can occur when using NTLM authentication
over TCP.
Images Affected:
- [SYSLIB]DCE$LIB_SHR.EXE
o Reduce virtual memory use by RPC runtime
Images Affected:
- [SYSLIB]DCE$LIB_SHR.EXE
o Increase socket connection queue length from 5 to 1024
Images Affected:
- [SYSLIB]DCE$LIB_SHR.EXE
o NTLM authentication errors show up as who_are_you failures.
Images Affected:
- [SYSLIB]DCE$LIB_SHR.EXE
o Attempting to start a DCE server with a well known endpoint,
such as RPCD (port 135) or PERF server (port 2001), fails with
sn "UNABLE TO BIND SOCKET" error, when there is no process
using the port.
Images Affected:
- [SYSLIB]DCE$LIB_SHR.EXE
o An OpenVMS DCOM client to a NT component fails with an RPC
"Unknown reject" error.
Images Affected:
- [SYSLIB]DCE$LIB_SHR.EXE
o DCE temporary files may appear in scratch or default
directories. If an error or control Y is entered during a
dce_setup invocation, temporary files, such as:
DCE_OUTnnnnnnnnn.TMP
are left in the directory pointed to by the logical
SYS$SCRATCH.
Images Affected:
- [SYSMGR]DCE$RPC_STARTUP.COM
o MFPR_xxx and MTPR_xxx PALcode instructions can leave registers
R1, R16 and R17 with unpredictable results. These registers
were not always saved and restored in ASTDEL_STACK.M64.
Although corruptions of these registers have not been known to
occur, the potential is there, particularly on newer
platforms.
Images Affected:
- [SYS$LDR]PROCESS_MANAGEMENT.EXE
- [SYS$LDR]PROCESS_MANAGEMENT.STB
- [SYS$LDR]PROCESS_MANAGEMENT_MON.EXE
- [SYS$LDR]PROCESS_MANAGEMENT.STB
o A DECthread may hang waiting for an event flag upcall.
Images Affected:
- [SYS$LDR]PROCESS_MANAGEMENT.EXE
- [SYS$LDR]PROCESS_MANAGEMENT.STB
- [SYS$LDR]PROCESS_MANAGEMENT_MON.EXE
- [SYS$LDR]PROCESS_MANAGEMENT.STB
o A call to $SETPRV may not result in the requested
modifications being applied to the security structures.
Subsequent calls requiring the unset privileges may fail with
SS$_NOPRIV. This only occurs when the SYSGEN parameter
ARB_SUPPORT equals 3.
Images Affected:
- [SYS$LDR]PROCESS_MANAGEMENT.EXE
- [SYS$LDR]PROCESS_MANAGEMENT.STB
- [SYS$LDR]PROCESS_MANAGEMENT_MON.EXE
- [SYS$LDR]PROCESS_MANAGEMENT.STB
o A BASIC application terminates abnormally with the
BAS$_PROLOSSOR, DEVFOREIGN or ACCVIO status.
Images Affected:
- [SYS$LDR]PROCESS_MANAGEMENT.EXE
- [SYS$LDR]PROCESS_MANAGEMENT.STB
- [SYS$LDR]PROCESS_MANAGEMENT_MON.EXE
- [SYS$LDR]PROCESS_MANAGEMENT.STB
o SET PROCESS/PRIORITY=n fails with the following error:
"%SET-E-NOTSET, need ALTPRI privilege to elevate above base priority"
when the target process is on a remote cluster node.
Images Affected:
- [SYS$LDR]SYS$CLUSTER.EXE
o Infrequent system crashes (INCONSTATE bug check in
SYS$FGEDRIVER) can occur while doing Fibre Channel I/O.
Images Affected:
- [SYS$LDR]SYS$FGEDRIVER.EXE
o When DECnet/OSI transport is run over IP, the DCL commands
MONITOR CLUSTER or MONITOR CLUSTER/NODE=xxxx result in the
following error messages:
%MONITOR-I-ESTABCON, establishing connection to remote node(s).
T_RCV: : system error - invalid buffer length
%MONITOR-E-COLLERR, error during data collection
-VPM-W-RECVERR, an error was encountered receiving remote data
Images Affected:
- SYS$COMMON:[SYSEXE]VPM.EXE
- SYS$COMMON:[SYSEXE]VPM_SERVER.EXE
- SYS$COMMON:[SYSLIB]SPISHR.EXE
- SYS$COMMON:[SYSLIB]MONTORSHR.EXE
- SYS$COMMON:[SYSEXE]MONITOR_TV.EXE
o SHOW DEVICE/FULL does not display density or compaction
supported and enabled/disabled accurately. The density
requested by the user is not propagated accurately for
MK-based devices, TMSCP served to clients.
When DEFAULT density is specified on the client, the density
returned from the server for MK-based devices is incorrectly
changed to 01XX codes. When displayed by SHOW
DEVICE/FULL, it displays as UNKNOWN density.
Images Affected:
- [SYSEXE]SET.EXE
o An unneccesary "Device queue full" status can be received,
even when the adapter is not busy.
Images Affected:
- [SYS$LDR]SYS$PKQDRIVER.EXE
o Failed transfers could incorrectly be considered successful,
resulting in possible data corruption problems.
Images Affected:
- [SYS$LDR]SYS$PKQDRIVER.EXE
o Error reporting in some error paths (port shutdown) are too
ambiguous to isolate the reasons for the shutdown.
Images Affected:
- [SYS$LDR]SYS$PKQDRIVER.EXE
o The PKQ port could come up offline, if there was not enough
BAP memory. The driver firmware would attempt to load, even
if the console firmware was already loaded and current.
Images Affected:
- [SYS$LDR]SYS$PKQDRIVER.EXE
o PKQDRIVER fails to load properly in memory systems larger than
2 GB.
Images Affected:
- [SYS$LDR]SYS$PKQDRIVER.EXE
o The RSX updates to ODS-1 date storage for Y2K compliance are
not compatible with the VMS sliding date implementation. This
correction updates the dump to interpret ODS-1 format dates in
the RSX implemented format.
Images Affected:
- [SYSEXE]DUMP.EXE
o If using the Text libraries instead of include paths to
resolve all the needed header files, the C compiler will not
read nested include statements greater than 16 deep from
within a text library.
Images Affected:
- [SYS$LDR]LBRSHR.EXE
o The routine LIB$CREATE_DIR would occasionally return an
improper error when the directory file that it was creating
was also being created, at the same time, by another process.
This would be seen as a false SS$_NOSUCHFILE error. The
directory file did exist after the call.
Images Affected:
- [SYSLIB]LIBRTL.EXE
- [SYSLIB]Update to STARLET.OLB
o LIB$FIND_VM_ZONE does not show all zones when there are
multiple zones created and the algorithm is VM_FIXED.
Images Affected:
- [SYSLIB]LIBRTL.EXE
- [SYSLIB]LIBRTL_INSTRUMENTED.EXE
o STR$COPY_R may erroneously return STR$_FATINTERR when freeing
the previous contents of a dynamic string. This problem may
also occur with the routine LIB$SCOPY_R_DX. Both routines are
frequently used by other RTL routines to return string values,
so, STR$_FATINTERR may be returned or signaled by a number of
RTL routines, including the BASIC RTL.
Images Affected:
- [SYSLIB]LIBRTL.EXE
- [SYSLIB]LIBRTL_INSTRUMENTED.EXE
o STR$DIVIDE produces incorrect results for certain divisions.
For example:
A: sign = 0, exp = -2, digits = "41291"
B: sign = 0, exp = -6, digits = "787564"
rndtrc = 1, totdig = 2
incorrect result was
C: sign = 0, exp = -2, digits = "28147068226765"
correct result is
C: sign = 0, exp = -2, digits = "52429"
Note that STR$DIVIDE is used by the basic RTL.
Also, STR$DIVIDE will report a memory allocation error, if the
TOT_DIGITS parameter is negative.
Images Affected:
- [SYSLIB]LIBRTL.EXE
- [SYSLIB]LIBRTL_INSTRUMENTED.EXE
o An INIT/MED=NOCOMP command appears to have worked properly
according to a SHOW DEV/FULL command executed after the INIT
command is issued. This will display COMPACTION = DISABLED,
when in fact it is ENABLED. This problem is only detected on
a visual inspection of the LEDS on the drive itself. They
indicate COMPACTION.
Images Affected:
- [SYSLIB]INIT$SHR.EXE
o An INIT/DEN=DEFAULT command to a local SCSI drive fails with
an IVDENS error.
Images Affected:
- [SYS$LDR]INIT$SHR.EXE
- [SYSEXE]INIT.EXE
o ODS-1 tapes yield dates with years of 00 in 2000; the "proper"
text string for year 2000 is ":0"
Images Affected:
- [SYSEXE]MTAAACP.EXE
o %MOUNT-F-MEDOFL and %MOUNT-F-FILACCERR errors occur when
attempting to mount the second volume of a two-volume RMU
backup.
Images Affected:
- [SYSEXE]MTAAACP.EXE
o The $Registry system service will return an error if
OPEN/CLOSE key operations are repeated a large number of
times.
The following sequence will produce REG$_INVKEYID,
REG$_INVPATH, REG$_NOKEY or REG$_NOVALUE errors upon
subsequent references to the registry key or its associated
value:
o Open and close the registry key 32k+ times.
o Exit the image.
o Open and close the registry key enough times to reach the
next multiple of 32k opens.
o Open and close the registry key 224k times without exiting
the image.
This does not require that the same key name be used each time.
It is, however, affected by the number of keys opened at a
given time and by the order in which keys are opened or
closed. Repeating the operation with the same number of keys
opened each time increases the likelyhood of incurring this
error.
The exact error returned will depend on the type of reference.
If you used the key ID with a subkey name, you will get a
REG$_INVKEYID or REG$_INVPATH error. If you use the key ID
alone, you will get a REG$_NOKEY or REG$_NOVALUE error.
The $Registry system service may allow too much process P1
pool to be allocated to Registry keys. This can occur if the
SYSGEN parameter CTLPAGES is set below 175 (the default is
256).
The $Registry system service intentionally returns an
SS$_INSFMEM error if you attempt to open so many keys that it
would not leave enough process P1 pool for other purposes.
This check was previously applied when $Registry needed to
expand its P1 pool allocation, but was not applied when
$Registry is first initializing, which also results in P1 pool
allocation. Since it is possible to set CTLPAGES below 175,
an additional check has been added to verify that the number
of remaining pages during initialization is at least 175 and
return an SS$_INSFMEM error if there are too few pages
remaining.
Images Affected:
- [SYS$LDR]SYS$NTA.EXE
o Per-thread security rights handling needs to be modified. The
modifications include adding sanity checks, re-writing some
modules for simplification and maintainability, eliminating
erroneous logic and ensuring a more consistent handling of
rights structures across routines.
Images Affected:
- [SYS$LDR]SECURITY.EXE
- [SYS$LDR]SECURITY_MON.EXE
o The F11CD routine READ_ATTRIBUTES does not properly return an
RSX-11 compatible ASCII date string. As per the RSX group
request, years beyond 1999 are processed in the ODS-1 ASCII
date format by continuing to increment the ASCII code of the
decades digit.
Images Affected:
- [SYSEXE]F11CACP.EXE
- [SYSEXE]F11DACP.EXE
o Provide the capability to pre-fault pages into memory without
waiting.
Images Affected:
- [SYS$LDR]SYS$VM.EXE
- [SYS$LDR]SYS$VM.STB
- [SYSLIB]SYS$PUBLIC_VECTORS.EXE
- [SYSLIB]SYS$SSISHR.EXE
o The macro defined in the PPP_LOG.C module for outputting
characters is not accounting for them. Characters that should
be counted are not counted.
Images Affected:
- [SYS$LDR]SYS$PPPDRIVER.EXE
o PPPP Connections fail.
Images Affected:
- [SYS$LDR]SYS$PPPDRIVER.EXE
o Modem signals do not drop when the line hangs up.
Images Affected:
- [SYS$LDR]SYS$ASNDRIVER.EXE
o When trying to write to a full AUDIT_SERVER mailbox, the
AUDIT_SERVER hangs in exit loop (IPL 0) and other processes on
the system hang in RWMBX state,
Images Affected:
- [SYSEXE]AUDIT_SERVER.EXE
o If the target initiated negotiations, the driver would reject
the negotiation attempt by sending a reject message. And if
it did not successfully complete WDTR negotiations, the driver
would reset to asynchronous data transfers.
Images Affected:
- [SYS$LDR]SYS$PKWDRIVER.EXE
o There are negotiation issues found with the KZPCA on the
PKWDRIVER. In particular, the KZPCA does not negotiate:
1. On a request sense or inquiry command to handle situations
in which a disk may have been hotswapped.
2. On ULTRA speed, even if the SCSI bus is single-ended, if
the adapter is ULTRA2-capable.
Images Affected:
- [SYS$LDR]SYS$PKWDRIVER.EXE
o Shutting down a node which shares cluster system files on
shadowed disks could result in a shadow merge operation at
reboot.
Images Affected:
- [SYSEXE]SHUTDOWN.COM
o An instruction in the code, which enables hardware
synchronization, was in the wrong place.
Images Affected:
- [SYSLIB]DECW$SERVER_DDX_GZ.EXE
o Using a combination of logical names and dynamic image
activation via LIB$FIND_IMAGE_SYMBOL, it is possible to
erroneously activate more than one copy of an image.
Images Affected:
- [SYS$LDR]IMAGE_MANAGEMENT.EXE
o There is a syntax error in the UINT32_PQ typedef statement,
which is contained in the FAR_POINTERS.H file. A semicolon
was omitted to terminate the statement, and during a C or C++
compilation, one of the following error messages displays:
typedef unsigned __int64 VOID_PQ; /* 64-bit pointer to arbitrary data*/
^
%CXX-E-EXPSEMICOLON, expected a ";"
at line number 107 in module FAR_POINTERS of text library
SYS$COMMON:[SYSLIB]SYS$STARLET_C.TLB;1
or
int ioc_std$q_internal_irp(IRP *irp, UCB *ucb, IOSB_PQ iosb, VOID_PQ mbz);
...............................................................^
%CXX-E-UNDECLARED, identifier "VOID_PQ" is undefined
at line number in module IO_ROUTINES of text library
SYS$COMMON:[SYSLIB]SYS$LIB_C.TLB;1
SYS$STARLET_C.TLB and SYS$LIB_C.TLB are text libraries either
containing FAR_POINTERS.H or referring to it and VOID_PQ is
the typedef statement on the line following the typedef with
the missing semicolon.
Images Affected:
- [SYSLIB]SYS$STARLET_C.TLB
o TBK$SHOW_TRACEBACK calls may cause ACCVIOs.
Images Affected:
- [SYS$LDR]TRACE.EXE
o When installing the ELSA Gloria Syergy Graphics card, the
driver returns a value that set the device as UNAVAILABLE.
Images Affected:
- [SYS$LDR]SYS$GZDRIVER.EXE
o The size of the direct DMA window on secondary GALAXY
instances was incorrectly returned.
Images Affected:
- [SYS$LDR]SYS$CPU_ROUTINES_2208.EXE
o The TurboLaser, when booted using the serial console, does not
configure the keyboard and mouse.
Images Affected:
- [SYS$LDR]SYS$GYCDRIVER.EXE
New Functionality Addressed in VMS721_DS20E_ES40-V0100:
o This kit provides OpenVMS Version 7.2-1 support for the Compaq
AlphaServer DS20E. The DS20E is a follow-on to the Compaq
AlphaServer DS20. This kit includes the following OpenVMS
support on the DS20E:
+ Server management
+ 3X-DEPVD-AA PCI adapter
+ IDE CD-ROM and boot and runtime
Image(s) Affected:
- [SYS$LDR]SYS$CPU_ROUTINES_2208.EXE
- [SYS$LDR]SYS$IIDRIVER.EXE
- [SYSEXE]SYS$SMHANDLER.EXE
o On Compaq AlphaServer DS20E systems, the following system routines
cannot be used to perform I/O tribyte reads and writes:
+ IOC$READ_PCI_CONFIG
+ IOC$WRITE_PCI_CONFIG
+ IOC$READ_IO
+ IOC$WRITE_IO
If a device driver calls any of these system routines with a
length of three, you must use one of the following methods
instead-depending on your I/O cards characteristics:
For IOC$READ_IO and IOC$READ_PCI_CONFIG:
+ Use a longword read, and mask out the byte.
+ Do a combination of word and byte reads, and
append the Data.
For IOC$WRITE_IO and IOC$WRITE_PCI_CONFIG:
+ Read a longword, modify the tribyte, and rewrite the
longword.
Note that AlphaServer 8200/8400 and GS60/140 systems with
Alpha 21264 CPUs support tribyte reads and writes.
o TCP/IP for OpenVMS Version 4.2 does not support the new
3X-DEPVD-AA PCI adapter on the AlphaServer DS20E. Prior
to running UCX$CONFIG procedures, use the following command
to manually configure TCP/IP with the new ethernet adapter:
_$ /INTERNET_INTERFACE=I /TYPE=(ETHERNET,CLUSTER) -
_$ /INTERNET_INTERFACE=I /TYPE=(ETHERNET,CLUSTER) -
_$ /DESCRIPTION=INTEL_ETHERNET
Check with your Compaq support representative if you have
further questions on TCP/IP and the 3X-DEPVD-AA PCI adapter.
o This change adds P2V chip support to the P2A code used on ELSA.
This functionality is only used if you are using a P2V graphics
adaptor
Image(s) Affected:
- [SYS$LDR]SYS$GZVDRIVER.EXE
- [SYSEXE]SYS$CONFIG.DAT
o This Section describes the requirements and procedures for
creating an OpenVMS Galaxy computing environment on an
AlphaServer ES40.
For complete information about the Galaxy Software Architecture
on OpenVMS Alpha, refer to the OpenVMS Alpha Galaxy Guide (Order
number: AA-REZQB-TE). The most recent version is always available
at
http://www.openvms.digital.com:8000/
For complete information about AlphaServer ES40 systems, see the
AlphaServer ES40 Owners Guide (Order number: EK-ES240-UG.A01).
Image(s) Affected:
- [SYSEXE]GCU.EXE
- [SYSMGR]GALAXY.GCR
- [SYS$LDR]SYS$CPU_ROUTINES_2208.EXE
- [SYS$LDR]SYS$CPU_ROUTINES_2508.EXE
- [SYSEXE]SYS$SMHANDLER.EXE
To create an OpenMVS Galaxy on an AlphaServer ES40 system:
1. Read the release notes chapter in the OpenVMS Alpha
Galaxy Guide.
2. Read the configuration and hardware requirements.
3. Perform the steps described below.
Before you start:
In addition to the information in the Release Notes chapter
in the OpenVMS Galaxy Guide, you must also be familiar with
the following configuration and hardware requirements:
+ Two-instance maximum
You can run a maximum of two instances of OpenVMS on an
AlphaServer ES40.
+ Console firmware
To create an OpenVMS Galaxy environment on AlphaServer
ES40 systems, you must download the latest version of the
V5.5-xx console firmware from the following location:
http://ftp.digital.com/pub/Digital/Alpha/firmware/interim/es40/
+ AlphaServer ES40 clock
An AlphaServer ES40 has one clock. For an OpenVMS Galaxy,
this means that you cannot run the two instances at
different times. Also, the SET TIME command affects both
instances. Note that this may not become evident until a
number of hours have passed.
+ Console ports
COM1 (lower) is the console port for instance 0.
COM2 (upper) is the console port for instance 1.
Unlike creating an OpenVMS Galaxy on an AlphaServer 8400,
you do not need additional hardware for the second
console. COM-2 is used for this purpose.
+ CPUs
CPU0 must be the primary for instance 0.
CPU1 must be the primary for instance 1.
CPUs 2 and 3 are optional secondary CPUs that can be
migrated.
+ I/O adapters
PCI Hose 0 (PCI0) belongs to instance 0 (upper 4 PCI slots)
PCI Hose 1 (PCI1) belongs to instance 1 (lower 6 PCI slots)
Note that PCI0 contains an embedded ISA controller.
+ Storage controllers
You will need one storage controller (such as a KZPSA) per
instance. For each instance, this can go to a separate
Storagework box or to the same box for running as a SCSI
cluster.
+ Network cards
If each instance needs network access, a network card
(such as a DE500) is required for each instance.
One card each goes in PCI0 and PCI1.
+ Physical memory
Private memory must start on a 64 MB boundary.
Shared memory must begin on an 8 MB boundary.
All instances except the last must have a multiple of
64 MB.
Step 1. Confirm the AlphaServer ES40 Configuration
Use the SHOW CONFIG command to make sure that the AlphaServer
ES40 you are using to create an OpenVMS Galaxy environment
meets the requirements described.
At the console prompt, enter the following command:
P00>>>show config
The console displays information similar to the following
example:
Firmware
ARC Console: v5.70
PALcode: OpenVMS PALcode V1.61-2
Tru64 UNIX PALcode V1.54-2
Serial Rom: V2.2-F
RMC Rom: V1.0
RMC Flash Rom: T2.0
Processors
CPU 0 Alpha 21264-4 500 MHz 4MB Bcache
CPU 1 Alpha 21264-4 500 MHz 4MB Bcache
CPU 2 Alpha 21264-4 500 MHz 4MB Bcache
CPU 3 Alpha 21264-4 500 MHz 4MB Bcache
Core Logic
Cchip DECchip 21272-CA Rev 9(C4)
Dchip DECchip 21272-DA Rev 2
Pchip 0 DECchip 21272-EA Rev 2
Pchip 1 DECchip 21272-EA Rev 2
TIG Rev 10
Memory
Array Size Base Address Intlv Mode
--------- ---------- ---------------- ----------
0 4096Mb 0000000000000000 2-Way
1 4096Mb 0000000100000000 2-Way
2 1024Mb 0000000200000000 2-Way
3 1024Mb 0000000240000000 2-Way
10240 MB of System Memory
Slot Option Hose 0, Bus 0, PCI
1 DAPCA-FA ATM622 MMF
2 DECchip 21152-AA Bridge to Bus 2, PCI
3 DEC PCI FDDI fwb0.0.0.3.0 00-00-F8-BD-C6-5C
4 DEC PowerStorm
7 Acer Labs M1543C Bridge to Bus 1, ISA
15 Acer Labs M1543C IDE dqa.0.0.15.0
dqb.0.1.15.0
dqa0.0.0.15.0 TOSHIBA CD-ROM XM-6302B
19 Acer Labs M1543C USB
Option Hose 0, Bus 1, ISA
Floppy dva0.0.0.1000.0
Slot Option Hose 0, Bus 2, PCI
0 NCR 53C875 pkd0.7.0.2000.0 SCSI Bus ID 7
1 NCR 53C875 pke0.7.0.2001.0 SCSI Bus ID 7
dke100.1.0.2001.0 RZ1BB-CS
dke200.2.0.2001.0 RZ1BB-CS
dke300.3.0.2001.0 RZ1CB-CS
dke400.4.0.2001.0 RZ1CB-CS
2 DE500-AA Network Con ewa0.0.0.2002.0 00-06-2B-00-0A-58
Slot Option Hose 1, Bus 0, PCI
1 NCR 53C895 pka0.7.0.1.1 SCSI Bus ID 7
dka100.1.0.1.1 RZ2CA-LA
dka300.3.0.1.1 RZ2CA-LA
2 Fore ATM 155/622 Ada
3 DEC PCI FDDI fwa0.0.0.3.1 00-00-F8-45-B2-CE
4 QLogic ISP10x0 pkb0.7.0.4.1 SCSI Bus ID 7
dkb100.1.0.4.1 HSZ50-AX
dkb101.1.0.4.1 HSZ50-AX
dkb200.2.0.4.1 HSZ50-AX
dkb201.2.0.4.1 HSZ50-AX
dkb202.2.0.4.1 HSZ50-AX
5 QLogic ISP10x0 pkc0.7.0.5.1 SCSI Bus ID 7
dkc100.1.0.5.1 RZ1CB-CS
dkc200.2.0.5.1 RZ1CB-CS
dkc300.3.0.5.1 RZ1CB-CS
dkc400.4.0.5.1 RZ1CB-CS
6 DECchip 21154-AA Bridge to Bus 2, PCI
Slot Option Hose 1, Bus 2, PCI
4 DE602-AA eia0.0.0.2004.1 00-08-C7-91-0A-AA
5 DE602-AA eib0.0.0.2005.1 00-08-C7-91-0A-AB
6 DE602-TA eic0.0.0.2006.1 00-08-C7-66-80-9E
7 DE602-TA eid0.0.0.2007.1 00-08-C7-66-80-5E
Step 2: Install OpenVMS Alpha Version 7.2-1
No special installation procedures are required to run OpenVMS
Galaxy software. Galaxy functionality is included in the base
operating system and can be enabled or disabled using the
console command and system parameter values described later in
this chapter.
If your AlphaServer ES40 is not part of a SCSI cluster, you
must install OpenVMS Version 7.2-1 on two system disks-one
disk for each instance.
If your AlphaServer ES40 is part of a SCSI cluster with a
cluster-common system disk, install OpenVMS Version 7.2-1 on
one system disk.
For more information about installing the OpenVMS Alpha
operating system, see the OpenVMS Alpha Version 7.2 Upgrade
and Installation Guide.
Step 3: Upgrade the Firmware
To upgrade the firmware, use one of the following procedures:
Copy the firmware file to MOM$SYSTEM on a MOP-enabled server
that is accessible to the AlphaServer ES40. Enter the
following commands on the console:
P00>>> boot -fl 0,0 ewa0 -fi {firmware filename}
UPD> update srm*
Or, use the following commands:
P00>>> BOOT -FLAGS 0,A0 cd_device_name
.
.
.
Bootfile: {firmware filename}
.
.
.
Set Environment Variables
Configure the primary console for instance 0.
CPU0 is the primary for instance 0.
CPU1 is the primary for instance 1.
For descriptions of the Galaxy environment variables and
common values for them, refer to Chapter 6 in the OpenVMS
Alpha Galaxy Guide.
The following example is for an AlphaServer ES40 with three
CPUs and 512MB of memory divided into 256MB + 192MB + 64MB.
P00>>> set lp_count 2
P00>>> set lp_cpu_mask0 1
P00>>> set lp_cpu_mask1 6
P00>>> set lp_io_mask0 10
P00>>> set lp_io_mask1 20
P00>>> set lp_mem_size0 10000000
P00>>> set lp_mem_size1 c000000
P00>>> set lp_shared_mem_size 4000000
P00>>> set auto_action halt
If you have four CPUs and you want to assign all secondary
CPUs to instance 1, the lp_cpu_mask1 variable will be E. If
you split the CPUs between both instances, CPU 0 must be the
primary for instance 0, and CPU 1 must be the primary CPU for
instance 1.
The following example shows LP_CPU_MASK values for secondary
CPU assignments with primary CPUs.
Assign secondary CPU 2 with primary CPU 0 and secondary CPU 3
with primary CPU 1.
>>>set lp_cpu_mask0 5
>>>set lp_cpu_mask1 A
CPU Selection LP_CPU_MASK
0(primary partition 0) 2^0 = 1
1(primary partition 1) 2^1 = 2
2(secondary) 2^2 = 4
3(secondary) 2^3 = 8
The mem_size variables depend on your configuration and how
you want to split it up.
lp_io_mask0 must be set to 1
lp_io_mask1 must be set to 2
You must set the console environment variable AUTO_ACTION to
HALT. This will ensure that the system does not boot and that
you will be able to enter the Galaxy command.
Step 5: Initialize the System and Start the Console Devices
1. Initialize the system and start the Galaxy firmware by
entering the following commands:
P00>>> init
P00>>> lpinit ! same as galaxy command
After the self-test completes, the Galaxy command will
start the console on instance 1.
Note that when the I/O bus is divided between the two
Galaxy partitions, the port letter of a device might
change. For example, a disk designated as DKC300 when
the AlphaServer ES40 is a single system could become
DKA300 when it is configured as partition 0 of the
OpenVMS Galaxy.
2. Configure the console for instance 1.
3. Configure the system root, boot device, and other related
variables.
The following example settings are from an OpenVMS
Engineering system. Change these variables to meet the
needs of your own environment.
Instance 0
P00>>> set boot_osflags 12,0
P00>>> set bootdef_dev dka0
P00>>> set boot_reset off !!! must be OFF !!!
P00>>> set ewa0_mode twisted
Instance 1
P01>>> set boot_osflags 11,0
P01>>> set bootdef_dev dkb200
P01>>> set boot_reset off !!! must be OFF !!!
P01>>> set ewa0_mode twisted
4. Boot instance 1 as follows:
P01>>> boot
Once instance 1 is booted, log in to the system account
and edit the SYS$SYSTEM:MODPARAMS.DAT file to include the
following line:
GALAXY=1
Confirm that the SCSNODE and SCSSYSTEMID SYSGEN parameters
are correct. Run AUTOGEN as follows to configure instance
1 as a Galaxy member, and leave the system halted:
$ @SYS$UPDATE:AUTOGEN GETDATA SHUTDOWN INITIAL
5. Boot instance 0 as follows:
P00>>> boot
Once instance 0 is booted, log in to the system account
and edit the SYS$SYSTEM:MODPARAMS.DAT file to include the
following line:
GALAXY=1
Confirm that the SCSNODE and SCSSYSTEMID SYSGEN parameters
are correct. Run AUTOGEN as follows to configure instance
0 as a Galaxy member, and leave the system halted:
$ @SYS$UPDATE:AUTOGEN GETDATA SHUTDOWN INITIAL
6. Prepare the Galaxy to come up automatically upon
initialization or power cycle of the system. Set the
AUTO_ACTION environment variable on both instances to
RESTART.
P00>>> set auto_action restart
P01>>> set auto_action restart
7. Initialize the Galaxy again by entering the following
commands at the primary console:
P00>>> init
When the console displays the following confirmation
prompt, type Y:
Do you REALLY want to reset all partitions? (Y/N)
Alternatively, you could power-cycle your system, and the
Galaxy with both instances should bootstrap automatically.
Problems Addressed in VMS721_DS20E_ES40:
o The EXTENDED_ID flag in the SYS$CONFIG.DAT file is used to
allow full 64-bit ID checking on PCI devices. This will be
used for LAN and ATM devices to ensure that the device is a
qualified device.
A problem in the LOOKUP_ID code causes the ID mask to be
extended to 64-bits the first time the flag is seen, and
it is never reduced back to 32-bits for non-extended IDs.
This causes a failure to find some disk devices at boot, and
prevents booting.
The user sees an error that a device on the boot command line
could not be found, and the system halts.
Image(s) Affected: [SYS$LDR]EXEC_INIT.EXE
o When the 680 handler issued the code 22Bh to SYS$SMHANDLER it
did not display a message.
Image(s) Affected: [SYSEXE]SYS$SMHANDLER.EXE
Problems Addressed in the VMS721_FIBRECHAN-V0200 Kit:
o This change updates the configuration file to allow the LP8000
version of the KGPSA adapter to be recognized by OpenVMS.
Before this change, only the LP7000 is recognized.
Images Affected: [SYSEXE]SYS$CONFIG.DAT
o OpenVMS hangs if a system is booted with a loopback connector
in place on an Emulex LP7000 or LP8000 adapter
Images Affected:
- [SYS$LDR]SYS$PGADRIVER.EXE
- [SYS$LDR]SYS$FGEDRIVER.EXE
o It is possible for data corruption to occur when removing a
cable (or disabling a port) between switches in a fabric with
cascaded switches. Because the link which is disrupted has
fabric-wide scope, it is impossible to predict which hosts or
devices might see the corruption. The pattern of corruption
seen has usually been misplaced data, meaning that a portion
of the data for a given I/O request has appeared in the
correct buffer but at the wrong offset in that buffer
Images Affected:
- [SYS$LDR]SYS$FGEDRIVER.EXE
- [SYS$LDR]SYS$PGADRIVER.EXE
- [SYS$LDR]FC$GLOBALS.STB
- [SYSLIB]FC$SDA.EXE
o When removing a cable (or disabling a port) between switches
in a fabric with cascaded switches, it is possible to hang a
system which is doing I/O to a device in the fabric. It is
actually individual I/O requests which get hung, the result of
the link not transitioning to the UP state correctly. This
can show up as the loss of one or more paths to a multi-path
device (depending on how many links/paths are affected). In
the worst case, all paths to a FibreChannel system disk will
be lost resulting in a system hang.
Images Affected:
- [SYS$LDR]SYS$FGEDRIVER.EXE
- [SYS$LDR]SYS$PGADRIVER.EXE
- [SYS$LDR]FC$GLOBALS.STB
- [SYSLIB]FC$SDA.EXE
o A system may hang during FibreChannel configuration when a
KGPSA is connected to a Pleiades-II switch.
Images Affected:
- [SYS$LDR]SYS$FGEDRIVER.EXE
- [SYS$LDR]SYS$PGADRIVER.EXE
- [SYS$LDR]FC$GLOBALS.STB
- [SYSLIB]FC$SDA.EXE
o When large I/O transfers are done under a heavy I/O load (e.g.,
running Backup or doing a shadow set copy), excessive Medium
Offline and Mount Verify messages are seen on a Turbolaser
system. Mount Verification is performed often enough to have
a serious impact on performance.
Images Affected:
- [SYS$LDR]SYS$FGEDRIVER.EXE
- [SYS$LDR]SYS$PGADRIVER.EXE
- [SYS$LDR]FC$GLOBALS.STB
- [SYSLIB]FC$SDA.EXE
o It is possible to cause a DOUBLEDEALO crash on a system when
removing a cable (or disabling a port) between switches in a
fabric with cascaded switches. This crash can occur even if
the system is idle at the time. The crash will only happen if
a particular LP7000 or LP8000 sees more than 10 other ports on
the fabric. The number of ports an adapter sees can be
determined as follows:
- There is always one NULL port
- There are 2 ports on the switch to which the adapter is
directly connected
- There is one port for each multi-bus HSG80 port connected
to the same fabric (not necessarily the same switch)
- There is one port for each other FibreChannel adapter
connected to the same fabric, even if these other adapters
are in the same system.
This same probelm can also cause an invalid exception system
crash.
Images Affected:
- [SYS$LDR]SYS$FGEDRIVER.EXE
- [SYS$LDR]SYS$PGADRIVER.EXE
- [SYS$LDR]FC$GLOBALS.STB
- [SYSLIB]FC$SDA.EXE
o An invalid exception system crash during boot can occur on a
system with an adapter which has failed initialization.
Images Affected:
- [SYS$LDR]SYS$FGEDRIVER.EXE
- [SYS$LDR]SYS$PGADRIVER.EXE
- [SYS$LDR]FC$GLOBALS.STB
- [SYSLIB]FC$SDA.EXE
Problems Addressed in the VMS721_FIBRECHAN-V0100 Kit:
o A system hang may occur after a FibreChannel switch unit is
removed or a FibreChannel switch port is disabled.
The problem was caused by not restarting the driver's Queue
Manager after setting STDT state to SS$_MEDOFL.
o A single-CPU system may hang or a CPUSPINTWAIT crash (SMP)
may occur when a heavy I/O load is running to FibreChannel
devices.
o A single-CPU system hang and an SMP-system CPUSPINWAIT crash
may occur when a QUEUE FULL is seen while a target is paused and
the target times out.
o An INCONSTATE crash may occur after a FibreChannel link disruption
such as rebooting or power cycling the FibreChannel switch.
o An INCONSTATE bugcheck or corruption may occur, most often on
Turbolaser, but potentially on any FibreChannel transfer which
requires the use of map registers.
Black Hole and Erase Pattern pages are not being mapped into
the Direct DMA window properly, and are not being mapped with
map registers when necessary.
o Machine check crashes can occur while running FibreChannel
hardware.
The FibreChannel adapter CSRs need to be mapped into PCI
memory space instead of PCI I/O space.
o Under heavy load, FC disks enter mount verification, then exit
mount verification after a short delay. This can reduce
performance, and disrupt the desired path selection.
Multipath polling is done with an ordered SCSI command. This
unnecessarily requires the issuing node to single-stream its
I/O to the device, and for all outstanding I/O from other
nodes to complete before the polling I/O completes. This
causes commands to timeout. The solution is to use an
unordered SCSI command for multipath polling.
Problems Addressed in VMS721_PTHREAD-V0100:
o A customer's application, when run with upcalls enabled, can die
with an ACCVIO. This is the result of a stack overflow while
running the application's exit handling routine.
o When a thread calls pthread_exit() the entire process is incorrectly
terminated by a %CMA-F-EXIT_THREAD error instead of simply
terminating the calling thread.
o Pagefault Completion Errors
Pagefault completions accumulate (as observed via the "pthread show
-u" command to SDA), with the corresponding waiting threads having
been, apparently, unblocked without consuming (i.e., freeing) the
completions. Also, sometimes threads are found waiting indefinitely
for completions (although this effect was never successfully
reproduced in-house), while other, mismatched completions remain
pending.
o Reduced performance and/or process threads hang.
Problems Addressed in VMS721_F11X-V0100:
o An XQPERR bugcheck in LOCKERS can occur when the retry limit
on the F11B$x lock is reached.
This problem can occur when the owner of the $x lock is
running at a high process process priority and a number of
processes that are in a clustered system are also trying to
validate this lock, but at a lower process priority.
Image(s) Affected - [SYS$LDR]F11BXQP.EXE
o After releasing the current process's IPL/Fork lock, a
system can crash with an SPLACQERR bugcheck
Image(s) Affected - [SYS$LDR]F11BXQP.EXE
o A directory file becomes "corrupt" and DUMP/DIRECTORY
displays a block similar to the following:
Virtual block number 3574 (00000DF6), 512 (0200) bytes
0000 Directory Entry:
0000 Size: 508
0002 Version limit: 32767
0004 Type: 0 (FID)
0005 Name count: 24
0006 Name: COSLR1201_01_JUPICC2.LIS
001E Version: 7859 FID: (40993,5,0)
0026 Version: 7858 FID: (40990,1,0)
002E Version: 7857 FID: (40988,3,0)
...
01E6 Version: 7802 FID: (40455,1,0)
01EE Version: 7801 FID: (40454,1,0)
01F6 Version: 32767 FID: (16744447,65535,0)
01FE End of records (-1)
The directory shuffle code creates the above erroneous
directory entry for the following reasons:
1. So that a new directory buffer will have a valid
structure (this allows VALIDATE_DIRBLK to write
the block to disk); and
2. The entry will be spotted as incorrect (via VERIFY)
if the system crashes in the middle of this shuffle.
After the directory block (with the erroneous directory entry)
is written to disk, the bad entry is removed. A subsequent
call to READ_BLOCK assumes that the block comes from the
buffer cache and not from disk. Under heavy load, this
assumption may not be true as the directory block may have
been kicked out of the cache.
Image(s) Affected - [SYS$LDR]F11BXQP.EXE
o XQP DELETE code accepts an FCB (File Control Block) off the
limbo queue if not IO$V_DELETE. This prevents the
invalidation of VIOC cache blocks as the result of a RENAME
operation. This causes a large amount of XQP (FCB) and VIOC
(CFCB) non-paged pool usage as well as XQPERR bugchecks.
Image(s) Affected - [SYS$LDR]F11BXQP.EXE
o Under the following circumstances,
1. A directory with multiple headers (e.g., from a large ACL)
is deleted on one node (A) in a cluster; and
2. the directory had been previously accessed on another node
(B) in the cluster,
The files created with the previously deleted headers in step
1 would show up on node B with the error:
%SYSTEM-F-NOSUCHFILE, no such file.
Image(s) Affected - [SYS$LDR]F11BXQP.EXE
New Functionality Contained in VMS721_GRAPHICS-V0300:
o This change adds P2V chip support to the P2A code used on
PowerStorm 4D10T. This functionality is only used if you
are using a P2V graphics adapter
Image(s) Affected:
- [SYS$LDR]SYS$GZVDRIVER.EXE
- [SYSEXE]SYS$CONFIG.DAT
Problems Addressed in VMS721_GRAPHICS-V0300:
o A reboot on a DPWS 5/500, 5/600 or 5/433 system with a
PowerStorm 4D10T card will crash. The system, must be shut
down, a console init done, and the booted.
Image(s) Affected:
- [SYS$LDR]SYS$GZDRIVER.EXE
- [SYS$LDR]SYS$GZVDRIVER.EXE
o The PowerStorm 4D10T card displays video "noise". This noise
is most visible on certain backgrounds, such as the root weave
pattern that is drawn on secondary heads.
Image(s) Affected:
- [SYS$LDR]SYS$GZDRIVER.EXE
- [SYS$LDR]SYS$GZVDRIVER.EXE
o When using a large window, the screen can become corrupted,
especially when moving large windows.
Image(s) Affected: [SYSLIB]DECW$SERVER_DDX_GZ.EXE
o When using text editing (DECW$NOTEPAD as an example), the
I-bar text cursor would corrupt text when moved with the
cursor keys.
Image(s) Affected: [SYSLIB]DECW$SERVER_DDX_GZ.EXE
o A number of problems have been reported on the TGA2:
1. The cursor would occasionally not display on any head
beyond the second screen when using 24-planes, and moving
the cursor between screens.
2. When switching screens, a stale cursor occasionally was
left behind on the last screen.
3. A warm restart after a halt or crash would bugcheck in
GYCDRIVER.
4. When switching screens quickly, the cursor would
occasionally be mis-positioned on the wrong screen.
5. DMA was not disabled for systems with > 1GB or when behind
a bridge.
Image(s) Affected:
- [SYS$LDR]SYS$GZVDRIVER.EXE
- [SYS$LDR]SYS$INDRIVER.EXE
Problems Addressed in VMS721_UPDATE-V0100:
o When a copy operation that interrupts a merge operation is
terminating, it finds that there are no members marked for the
merge and the thread crashes the system with a SHADDETINCON
bug check.
Images Affected:
- [SYS$LDR]SYS$SHDRIVER.EXE
- [SYSEXE]SHADOW_SERVER.EXE
o SHOW DEVICES shows zero percent merged status although the
shadow set status does not indicate that a merge is required.
Images Affected:
- [SYS$LDR]SYS$SHDRIVER.EXE
- [SYSEXE]SHADOW_SERVER.EXE
o Bit 16 in SHADOW_SYS_DISK can be set by the user to eliminate
using remote members of the shadowset for reads. Occasionally,
use of bit 16 fails to eliminate remote members from being used.
Images Affected:
- [SYS$LDR]SYS$SHDRIVER.EXE
- [SYSEXE]SHADOW_SERVER.EXE
o A CPUSPINWAIT bug check can occur if the read of the SCB of
a shadow set member cannot pass the checksum test.
Images Affected:
- [SYS$LDR]SYS$SHDRIVER.EXE
- [SYSEXE]SHADOW_SERVER.EXE
o DCD (Disk Copy Data) will not always be initiated properly.
During an assisted copy operation, if the source member was
dismounted or otherwise removed from the shadow set, the
connection to the controller would not clean up correctly.
Images Affected:
- [SYS$LDR]SYS$SHDRIVER.EXE
- [SYSEXE]SHADOW_SERVER.EXE
o A full copy operation that is interrupted for a mini-merge
will not complete the full copy operation correctly.
Images Affected:
- [SYS$LDR]SYS$SHDRIVER.EXE
- [SYSEXE]SHADOW_SERVER.EXE
o Typing incorrect commands results in a system crash.
Images Affected:
- [SYS$LDR]SYS$SHDRIVER.EXE
- [SYSEXE]SHADOW_SERVER.EXE
o When two disks are added to a shadowset in the same mount
command, the copies are done sequentially instead of in
parallel. This causes the copies to take twice as long as
they should.
Images Affected:
- [SYS$LDR]SYS$SHDRIVER.EXE
- [SYSEXE]SHADOW_SERVER.EXE
INSTALLATION NOTES:
This kit requires a system reboot. Compaq strongly recommends that
a reboot be performed immediately after kit installation to avoid
system instability. If there are other nodes in the VMScluster, they must
also be rebooted in order to make use of the new image(s).
If it is not possible or convenient to reboot the entire cluster at
this time, a rolling re-boot may be performed.
All trademarks are the property of their respective owners.
This patch can be found at any of these sites:
Colorado Site
Georgia Site
Files on this server are as follows:
dec-axpvms-vms721_update-v0100--4.README
dec-axpvms-vms721_update-v0100--4.CHKSUM
dec-axpvms-vms721_update-v0100--4.pcsi-dcx_axpexe
vms721_update-v0100.CVRLET_TXT
|
