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

• Allows hwmgr -redirect to work with lockmode 4.
• Fixes invalid queue entries and system panics following Retries Exhausted.
• Fixes a persistent reduction in SCSI queue depth after a queue full occurs.
• Fixes a race condition in xpt_async_event_thread.
• Fixes an “mcs_lock:lock already owned by cpu” panic.
• Fixes an issue with the storage subsystem I/O Barrier error handling.
• Fixes USB problems that prevent some USB keyboards and mice from working.
• Fixes POx Uncorrected and Unrecoverable Error panics on EV7 based systems.
• Corrects tag and non-tag I/O handling when the system is booting.
• Corrects a problem in which a multi-threaded application accessing a single tape drive via
a single FILE object shared amongst the threads causes a kernel memory fault.
• Fixes a problem with VMAC not working on the bcm driver (DEGXA) without explicitly
enabling promiscuous mode.
• Fixes a rare problem in which a “queue full” response during device discovery results in a
kernel memory fault.
• Adds additional fast fail functionality to the disk class driver.
• Clarifies the binary error-log entry for medium or hardware errors.
• Correct error logging problems in the emx driver.
• Fixes a problem that causes an “invalid link address” kernel panic in the Alteon gigabit
Ethernet driver.
• Fixes the incorrect locking order in alt driver.
• Corrects a problem in which FibreChannel dropped data is not detected during I/O error
recovery and so is not retried, thereby causing unexpected and unnecessary errors.
• Corrects a problem in which tape devices with no ddr entry would incorrectly enable or
disable compression to the device special file that was used.
• Fixes a problem in which the operating system receives an environmental machine check
packet from the firmware, but fails to correctly recognize the sensor that is identified as
faulty by the machine check.
• Corrects a problem in which multiple access to changer devices (for example, via
Legato/Networker) could lead to a kernel memory fault in the changer driver.
• Fixes problems with DS15 environmental monitoring.
• Fixes mass storage disk subsystem handling of specific important events reported by the
HSG80 and HSG60 array controllers.
• Corrects a problem in which multiple processes accessing a changer device can hang in a
deadlock situation accessing the changer.
• Improves logging of sense data during error recovery.
• Fixes a problem with stalled or hung I/O's on clusters.
• Corrects the panic, “cmn_err: CE_PANIC: elan0: heartbeat pacemaker - flatlined ....” that
occurs on AlphaServer SC systems.
• Fixes a panic that can occur when the ee driver momentarily drops its lock.
• Resolves a logic error in hardware-specific platform code for certain boundary conditions.
• Enables IPv6 MTU to be set up to 9000 bytes on gigabit Ethernet.
• Fixes a rare kernel memory fault that occurs when booting an AlphaServer ES80.
• Fixes a timer queue overflow panic with ccmn_get_dbuf.
• Fixes a kernel memory fault in hwc_lookup_devt_safe.
• Enables the right mouse button for some two-button USB mice.
• Fixes a kernel memory fault that can occur during boot on some large AlphaServer GS1280
systems.
• Enhances CAM disk driver resiliency.
3.4 Summary of Base Operating System Patches 123