HP Tru64 UNIX and TruCluster Server Version 5.1.B-4 Patch Summary and Release Notes (13156)

• Corrects a problem that causes a system panic while running applications that open a RAID
device and the faulting routine is control_port_open.
• Fixes an I/O hang condition on Fibre Channel.
• Fixes a rare case in which the target of a ladebug-invoked routine (via the call func command)
aborts with a segmentation fault while derefencing the gp register.
• Corrects a problem in which an unconfigured PCI devices can cause a panic during boot.
• Corrects a problem in which Tru64 UNIX sees an HP-XP RAID array controller as a disk
after an HP-XP storage device is added to the system.
• Corrects an address problem with hwmgr delete while a SCSI scan is in progress.
• Updates the Radeon driver to not reject multiheaded configurations.
• Removes an unwanted test text message.
• Fixes a memory fault condition in the emx driver that occurs when responding to an inquiry
command from a remote port in the fabric.
• Corrects a problem that reduces performance or causes a device to switch controllers
unnecessarily when active paths are improperly treated as standby or vice versa.
• Fixes a problems in which locks used for disk I/O are not released during rare error
conditions.
• Fixes a problem with disk I/O barrier code that was not executed when it was needed.
• Corrects a condition in which a panic can occur when unrecognized PCI devices are probed
with a PCI class code of 255. The problem is probably limited to very old or specialized
devices.
• Corrects a HSZ70 controller failover failure that generates the message "Logical unit not
ready, cause not reportable."
• Changes the way reservation-conflict errors are handled in a cluster.
• Prevents a potential panic that generates the message “memory_test=partial” or
“memory_test=none.”
• Corrects a problem in which certain older tape drives (such as the TZK50 and TSV07) would
produce only errors when they were accessed.
• Causes the AlphaServer GS1280 to accurately report and handle fatal memory errors if the
associated CPU is off line.
• Extends the hardware managers for CPU entries to include the attributes on a GS1280
platform to include the CPU chip revision number and the cache size.
• Fixes a condition that causes hung I/Os in systems with multiple RADs.
• Updates the tu driver to support Ethernet Multicast addresses larger than 512, which is
necessary to support a large number of IPv6 addresses
• Corrects a problem that causes the command sysconfig -r hwc hwc_print=5 to generate a
kernel memory fault.
• Updates the alt driver to V2.0.20 to fix issues with IPv6, NFS performance, and SNMP.
• Fixes an improper handling of domain, area, and fabric RSCNs (Registered State Change
Notifications) by the emx driver that results in the nondetection of path failures to storage
devices in the fabric."
• Fixes the reporting of device monitoring events and hardware errors during disk recovery
from the disk driver to the binary errlog.
• Corrects a problem that occurs with a device or bus reset during the execution of a command
to a media changer device, like a tape library, and causes the system to panic with the
message "PWS_CCB_QUE_REMOVE: ccb not on any list."
• Prevents a cluster panic when installing new RAID hardware before OS support is installed
on the cluster.
• Fixes a problem in which an I/O operation is prematurely treated as complete, which can
cause a panic in XPT or a storage driver.
3.4 Summary of Base Operating System Patches 121