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

• Fixes a problem with the USB keyboard driver that affects keyboard operation when the X
server is not running. Specifically, the autorepeat function it may appear jumpy, or keys
may appear to press themselves after they are released.
• Fixes a problem relating to gh_chunks allocation on some configurations.
• Fixes various problems in the bcm driver for DEGXA Gigabit Ethernet adapters.
• Provides the logging of an informational event to the errorlog when the peripheral driver
is unable to get a user specified IDENTIFIER for a storage array device.
• Fixes CAM errors that occur when opening CD devices containing blank media.
• Corrects a problem in which a read or write operation to a changer device creates an
unkillable process.
• Fixes a problem that can cause a kernel memory fault during boot if a Fibre Channel LSM
boot disk is not discovered during the device probe.
• Suppresses an in_cksum() console debug message.
• Fixes a memory fault condition in the user agent driver that occurs when multiple threads
issue I/Os on different CPUs.
• Fixes jitter problems in 24-bit depth on VX1 graphics cards with certain date codes.
• Fixes the cause of a system crash that can occur if the hwmgr command deletes a disk while
it is in recovery.
• Increases BOOTP error timeout values for RIS kernel installations.
• Updates the values in the header of the PCI subpackets.
• Corrects problems that can result in changer devices not responding or in a kernel memory
fault when the FNDELAY flag is set in changer open and changer ioctl.
• Allows multiple PCI-X devices on a system to use Message Signaled Interrupts (MSI).
• Reduces the existing limitation on the number of CPUs that can be taken off line by allowing
the other CPUs in the system to handle the interrupts coming from up to two IO7s.
• Corrects a "simple_lock: time limit exceeded" issue that can occur in the CAM I/O subsystem
when MCS locking is disabled.
• Fixes a problem that turns off the reporting of correctable errors forever on any CPU, except
CPU 0, once the throttling of correctable errors has begun.
• Causes the posting of environmental events in the event of a machine check on EV7-based
platforms.
• Corrects a problem on AlphaServer ES45 platforms that resets sysconfigtab settings for the
attribute titan_sys_ps_hotswap after a system reboot.
• Provides modifications for possible future emx hardware.
• Provides support for the Philips USB controller, which is shipped on some AlphaServer
GS1280 systems. Without this patch, the Philips USB controller may fail to detect and
configure devices below it.
• Fixes various problems in the ee driver for DE60x Ethernet adapters.
• Corrects a problem in which /sbin/ddr_config does not accept values for ReadyTimeSeconds
larger than 255. The new limit is 86400 seconds (24 hours).
• Fixes problems with NUMA disk statistics.
• Fixes a KMF problem can occur if some nodes in cluster are rebooted and a device is shared
by all the nodes.
• Changes the CAM subsystem message that is printed to the error log on a recovered read
error from "bad block number" to "block number."
• Provides hardware support for the AlphaServer DS15/TS15 platform.
• Corrects a race condition that can cause a process to hang during disk error recovery failure
processing.
• Fixes a hole in disk I/O handling that can cause a panic in rare circumstances.
120 Tru64 UNIX Patches