Red Hat Training

A Red Hat training course is available for Red Hat Enterprise Linux

8. Resolved Issues

8.1. All Architectures

  • Netapp devices can now complete failback (after a previously-failed path is restored) within a reasonable time with the default dm-multipath configuration.
  • system-config-kickstart now supports package selection through the Red Hat Network plugin.
  • kudzu can now properly parse ifcfg-* files that contain quotes around the HWADDR or SUBCHANNELS parameters. In addition, if a device's HWADDR is not specified, kudzu no longer modifies network configurations when the device is changed.
  • Running netstat with the -A inet or -A inet6 option (on a system where sctp was not added to the kernel) no longer terminates abnormally. Note, however, that netstat will display the following warning message when invoked with the -s option:
    	      netstat: no support for `AF INET (sctp)' on this system.
    
  • The nohide export option is no longer required on referral exports (i.e. exports that specify a referral server). For more information on bound mounts, refer to man 5 exports.
  • The priority callouts of dm-multipath are now statically compiled and copied onto the memory of multipathd. As such, multipathd no longer requires access to the root file system in order to execute priority callouts.
    This fixes a problem that occurred when running dm-multipath on devices containing the root file system, which caused such devices to freeze during fibre-channel path faults.
  • parted can now understand and correctly print out Xen Virtual Device (XVD) partition labels. This allows paravirtualized guests to use the parted utility. It is no longer necessary to run parted within dom0 to configure disk partitions on paravirtualized guests.
  • When upgrading to Red Hat Enterprise Linux 5.2 via Red Hat Network, you no longer need to manually import the redhat-beta key prior to upgrading.
  • Systems using the AMD 8132 or HT 1000 chipsets no longer need the pci=nommconf kernel parameter.
    Note that the system will still restrict such bridges to using the PortIO CF8/CFC mechanism. However, bridges (including those on the same platform) that respond correctly to MMCONFIG cycles will use MMCONFIG, provided that the platform's BIOS correctly supports MMCONFIG.
    However, if you are installing Red Hat Enterprise Linux 5.2 on an HP DC7800, the kernel parameter, pci=nommconf, is still needed. This is because the HP DC7800 is not yet included in the MMCONF blacklist.
  • Previous versions of Red Hat Enterprise Linux 5 on HP BL860c blade systems could hang during the IP information request stage of installation. When this occurred, you were required to reboot and perform the installation with Ethernet autonegotiation disabled.
    Red Hat Enterprise Linux 5.2 includes a fix for this issue.
  • Gemalto 64K smart cards now use Chip/Smart Card Interface Devices-compliant (CCID-compliant) readers. Previously, this smart card used the built-in e-gate reader, which essentially meant that the card and reader were being inserted at the same time. As a result, coolkey did not consistently recognize Gemalto 64K smart cards.
    With this update, coolkey now works correctly with Gemalto 64k smart cards.

8.2. x86-64 Architectures

  • Systems equipped with Intel graphics cards that have the XD (execute disable) option disabled no longer crash upon exiting an X session.
  • Using the divider= command-line argument now works properly on this architecture.
  • The Calgary IOMMU chip is now supported by default. As such, the kernel command line option iommu=calgary to enable support for this chip is no longer necessary.

8.3. ia64 Architecture

  • Kernel modules such as e1000 and qla2xxx can now be unloaded while running the virtualized kernel. As such, you no longer need to reboot the system after installing third-party drivers.
  • The X server no longer attempts to utilize memory regions incompatible to its needs. This fixes a bug that previously caused a Machine Check Abort (MCA) on some Itanium systems.
  • A machine check error (which previously resulted in a dom0 reboot) no longer occurs when you attempt to run the virtualized kernel on systems that use the P600 SmartArray controller.