Source

RETAIN tip: H193124

Symptom

The system experiences any combination of the following issues:

• Front side bus faults
• Machine check errors
• Kernel panics

The details of the error messages can differ greatly and may occur at any time when the operating system is up. Machines will shut down and/or reboot after the failure.

These failures are only exhibited on 64-bit versions of Red Hat 4 and Red Hat 3 (x86_64 kernels). 32-bit versions are not affected (i386 kernels).

To verify whether the the CPU is affected, run the following command from the terminal:

cat /proc/cpuinfo

An affected Intel CPU will show more than 36 bits of physical addressing, such as the following:

address sizes: 38 bits physical, 48 bits virtual

An affected AMD CPU will show more than 40 bits of physical addressing, such as the following:

address sizes: 48 bits physical, 48 bits virtual

Affected configurations

The system may be any of the following IBM servers:

• BladeCenter HS12, type 8014, any model
• BladeCenter HS12, type 8028, any model
• BladeCenter HS21 XM, type 1915, any model
• BladeCenter HS21 XM, type 7995, any model
• BladeCenter HS21, type 1885, any model
• BladeCenter HS21, type 8853, any model
• BladeCenter LS21, type 7971, any model
• BladeCenter LS41, type 7972, any model
• IntelliStation Z Pro, type 9228, any model
• IntelliStation Z Pro, type 9232, any model
• System x3400, type 7973, any model
• System x3400, type 7974, any model
• System x3400, type 7975, any model
• System x3400, type 7976, any model
• System x3500, type 7977, any model
• System x3550, type 1913, any model
• System x3550, type 7978, any model
• System x3610, type 7942, any model
• System x3650, type 1914, any model
• System x3650, type 7979, any model

This tip is not option specific.

The system is configured with at least one of the following:

• Red Hat Enterprise Linux 3, any update
• Red Hat Enterprise Linux 4 Update 1, Update 2, Update 3, Update 4, Update 5

Note: This does not imply that the network operating system will work under all combinations of hardware and software.

Please see the compatibility page for more information:

• IBM ServerProven

Solution

If running RHEL4:

This issue is resolved in applying Update 6.

If running RHEL3:

This issue is resolved by migrating over to RHEL4, Update 6, or contact Red Hat directly.

The referenced files are available from the Red Hat support web site at the following URL:

• http://www.redhat.com

Additional information

Intel CPUs have had 36-bit memory addressing and AMD CPUs have had 40-bit memory addressing for many years. Both vendors increased their memory addressing architecture in recent CPUs.

The kernel does not properly address the extra addressing bits in the CPUs, potentially causing memory corruption within the kernel. The updated kernel is corrected to properly use these address bits.

Below is Red Hat documentation on this issue:

• https://rhn.redhat.com/errata/RHSA-2007-0939.html
• http://kbase.redhat.com/faq/FAQ_42_11695.shtm
• http://kbase.redhat.com/faq/FAQ_85_11696.shtm