Intel E5345 Specification Update - Page 20

AJ10. Code Segment Limit/Canonical Faults on RSM May be Serviced before Higher Priority Interrupts/Exceptions and May Push the Wrong Address Onto the Stack Problem: Normally, when the processor encounters a Segment Limit or Canonical Fault due to code execution, a #GP (General Protection Exception) fault is generated after all higher priority Interrupts and exceptions are serviced. Due to this erratum, if RSM (Resume from System Management Mode) returns to execution flow that results in a Code Segment Limit or Canonical Fault, the #GP fault may be serviced before a higher priority Interrupt or Exception (e.g. NMI (Non-Maskable Interrupt), Debug break(#DB), Machine Check (#MC), etc.). If the RSM attempts to return to a non-canonical address, the address pushed onto the stack for this #GP fault may not match the non-canonical address that caused the fault. Implication: Operating systems may observe a #GP fault being serviced before higher priority Interrupts and Exceptions. Intel has not observed this erratum on any commercially available software. Workaround: None Identified. Status: For the steppings affected, see the Summary Tables of Changes. AJ11. A Write to an APIC Register Sometimes May Appear to Have Not Occurred Problem: With respect to the retirement of instructions, stores to the uncacheable memorybased APIC register space are handled in a non-synchronized way. For example if an instruction that masks the interrupt flag, e.g. CLI, is executed soon after an uncacheable write to the Task Priority Register (TPR) that lowers the APIC priority, the interrupt masking operation may take effect before the actual priority has been lowered. This may cause interrupts whose priority is lower than the initial TPR, but higher than the final TPR, to not be serviced until the interrupt enabled flag is finally set, i.e. by STI instruction. Interrupts will remain pending and are not lost. Implication: In this example the processor may allow interrupts to be accepted but may delay their service. Workaround: This non-synchronization can be avoided by issuing an APIC register read after the APIC register write. This will force the store to the APIC register before any subsequent instructions are executed. No commercial operating system is known to be impacted by this erratum. Status: For the steppings affected, see the Summary Tables of Changes. AJ12. Programming the Digital Thermal Sensor (DTS) Threshold May Cause Unexpected Thermal Interrupts Problem: Software can enable DTS thermal interrupts by programming the thermal threshold and setting the respective thermal interrupt enable bit. When programming DTS value, the previous DTS threshold may be crossed. This will generate an unexpected thermal interrupt. Implication: Software may observe an unexpected thermal interrupt occur after reprogramming the thermal threshold. Workaround: In the ACPI/OS implement a workaround by temporarily disabling the DTS threshold interrupt before updating the DTS threshold value. Status: For the steppings affected, see the Summary Tables of Changes. AJ13. Problem: Count Value for Performance-Monitoring Counter PMH_PAGE_WALK May be Incorrect Performance-Monitoring Counter PMH_PAGE_WALK is used to count the number of page walks resulting from Data Translation Look-Aside Buffer (DTLB) and Instruction Translation Look-Aside (ITLB) misses. Under certain conditions, this counter may be incorrect. 20 Intel® Xeon® Processor 5300 Series Specification Update, December 2010